by Dr. Laurence Goldman

The Huli are one of the largest cultural groups in the Southern Highlands Province of Papua New Guinea, and now number over 80,000. They inhabit an area of approximately 2000 square miles, and in the north there is an uninterrupted flow of population between Huli and Duna country; most people around Tanggi (see Map 1) are bi-lingual. The land is administratively divided between the Koroba sub-district and the Tari sub-district. The first Europeans to contact the Huli were Jack Hides (1936) and Peter O’Malley in the early 1930’s subsequent to which patrols by Sinclair (19SS,1957,1966,1973) and others resulted in the establishment of Government stations at Tari (in 1951) and Koroba (19SS). Extensive administration contact over the last three decades has wrought far-reaching social and economic changes in the culture. Cessation of warfare and a widespread network of roads has considerably increased the mobility of the people. It is not uncommon to find Huli in Port Moresby or Hagen. Economic development has been slow in comparison with other major Highland populations, but includes cattle farming, cash-cropping, coffee plantations and silk-worm projects. Traditional trade routes varied according to location and generalisation for the Huli area is an extremely dangerous venture. In the Koroba district, black-palm bows and Paradise feathers came largely from the Duna, salt from Enga, Tigasso tree-oil from peoples west of Levani who are linguistically related to the Tinali. Nowadays it is common for Kaluli (see Schieffelin 1976) to trade bows, string-bags, oil and Hornbill beaks with Huli as far north as Tanggi. While much of the traditional belief system remains intact, the larger ritual practices of Tege (fertility cycle), Dindi Pongo
(earth ritual) and Haroli (bachelor cult) died out in the early 1960’s. Most areas in Huli are now serviced by local aid-posts and schools, with the majority of people nominal converts to various introduced religions (e.g. Catholic, Wesleyan and Seventh-Day Adventists).

MAP 1: Koroba Sub-District, Southern Highlands Province

Development has not been uniform over the whole area and change is most evident the closer one is in proximity to a Government or Health centre. Money (£1 = 1.3Kina) is now generally accepted as a substitute for the traditional mediums of pigs or cowrie-shells, and with the imposition of a multiplicity of new legal institutions, the old authority structure has been considerably undermined. A rapid growth in population over the last twenty years, coupled with an increasing awareness of the economic potentialities of land resources, has resulted in an increase in land disputes. Nevertheless, despite all these changes it is still common for most Huli adult men to wear their wigs, and in outlying communities like Ialuba (see Map.1. ) where I conducted my fieldwork, western clothes were the exception rather than the norm. The origins of the Huli are obscure, though culturally and linguistically they are closest to the
Duna and Ipili Paiela (Enga}. Indeed, their own myths stipulate common origins for these three peoples. The area lies at an altitude of between 1500 and 2100 metres above sea-level. Rainfall averages 2500 millimetres annually, while temperatures vary between 7° C minimum and 27°c maximum. The topography varies from flat grasslands, to hilly and precipitous terrain. One finds sinkholes, escarpments, volcanic ranges and limestone outcrops, with dense rain forests giving way to small valleys. The Huli are subsistence farmers utilising a bush-fallow technique devoted to sweet-potato cultivation. This is supplemented by taro, sugar-cane, pandanus, and various leaf vegetables. The system of horticulture includes ring-barking and burning (e hiraya), composting (o dabia), fallowing, mounding (mondo) and drainage. They differentiate between three types of gardens; those made in the swamp (lara), those made immediately around the house (gama) and the larger mound tracts (mabu). Traditional tools were the wooden digging stick and stone axe, though now spades and steel axes are widely used. With a lack of distinguishable seasons there is little variation in the economic activities of the Huli, and certainly no annual horticultural cycle. In most areas small “markets” operate once a week, and are occasions on which people meet to sell produce, discuss news and pursue claims or grievances. Hunting for marsupials, snakes, cassowary and Bird of Paradise is more common in border areas like Ialuba, but is not of great dietary significance. The Huli settlement pattern is one of scattered households within named territorial units. The male-female opposition is here more marked than in most other Highland societies, with separate residential, eating and social arrangements. The first systematic work on Huli language was carried out by M. Rule (1954) ,with further research by B. Cheetham (1977a,1977b) adding considerably to our understanding of the system. Huli has been classified as a ember of the Enga-Huli-Pole-Wiru family (Wurm 1961). As will become clear from Chapter 2 onwards, Huli has three distinct tones but the status and significance of these is far from clear. I consider it a secondary feature of the language and I show that tonal oppositions do not always mark significant changes in the semantic import of otherwise similar words. Much of the minutiae of Huli life is given in the body of this thesis and I have here forsaken a lengthy outline of the culture. The Huli are well-known in the anthropological literature through the considerable publications of R. Glasse (1959a, 1959b, 1965, 1968), J. Pugh-Kitingan (1975, 1977, 1979) and myself (1979,1980).

Any ethnographer that follows after previous eminent anthroplogists finds himself inextricably fulfilling a number of functions with respect to the previous fieldwork. This can range from corroboration of ethnographic fact and continuation of theoretical advances in the interpretation of that culture, to utilising the work as a basis for study of social change. It would have been less than honest of me not to have been forthright in stating the considerable differences that exist between my data and that cited by Glasse in his accounts. Whilst my respect for the latter is in no way diminished by this situation, I have expressed the opinion that this hiatus is due mainly to misunderstandings about language and language use. In this context, I have felt it imperative to include, wherever possible, native language terms as a firm foundation for my own arguments, as well as future debate on possible areas of disagreement. With particular reference to kinship and descent, the divergences were of a great enough magnitude for me to have devoted considerable attention to the system. However, rather than presenting the analysis in a unified fashion, I have deliberately chosen to integrate my treatment of certain aspects with the case material as it is presented throughout the chapters. As stated previously, generalisation is dangerous, and the bulk of my data is relevant for the Koroba area only, though the structural patterns I elicit prevail throughout the Hult. The present is an historical one and should be understood to refer to conditions found during the period of fieldwork.

Kinship and the Male – Female Relationship

It is not enough to know what to say, we must also say it in the right way. Aristotle (Rhetoric xii)

The enduring consideration of the Huli as an anomalous case within the domain of New Guinea Highlands ethnography, has validated an early observation by J. Barnes that “Huli institutions are likely to remain analytically controversial” (1968:4). Comparative studies concerned with establishing valid cross-cultural generalisations, confronted the available data in diverse ways. The Huli were labelled A “test case”(Allen 1967:47) or designated atypical with respect to the combination of rituals for deities and the dead (cf. Lawrence & Meggitt 1965:26), the use of descent-constructs to determine recruitment modes (A. Strathern 1972:186), and a non-unilineal ideology at variance with patrilineal-patrivirilocal patterns prevalent in Highland societies (cf.De Lepervanche 1968:186). In contrast to those analyses where reference to the Huli was judiciously omitted (cf.A.Stratbern 1973, La Fontaine 1973), other studies abstracted random cultural factors in support of typologies and hypotheses focused on warfare (cf.Berndt
1964), male initiation cults (Allen 1967), and male-female relations (Meggitt 1964a). While inconsistencies in the early ethnography (ref.De Lepervanche 1968:187) were in part responsible for the enigmatic status of the Huli, development of new ethnographic and theoretical frameworks in the intervening years has rendered a re-examination of this culture expedient. The proliferation of studies in the last two decades dealing with aspects of descent and locality, engendered questions the answers to which could not alone be gleaned from the available data (i.e.Glasse 1959a, 1959b, 1965, 1968). More specifically, De Lepervanche (1968:186) – and later A.Strathern (1972:186) – remarked that the kinship system of the Huli displayed a number of morphological correspondences with the models of Choiseul and Kwaio societies, not shared by other Highland systems. I have attempted to locate the following analysis of Huli social structure within the framework of topics outlined above. As a preface to the argument developed, I consider it important to briefly state the theoretical precepts which structure this chapter.

A central feature of the analyses of Kwaio and Choiseul systems ,and indeed of such Highland ethnographies as the Melpa (Strathern 1972), is the separation of the realm of ideology from social process (Scheffler 1965:viii) – of the cognitive from the statistical (Keesing 1971:126). It is argued that description of social structures should keep distinct categories from groups, and the idioms used to refer to these from norms relating to recruitment processes. The concept of “membership” must therefore similarly be delineated by reference to both social categories and de facto groups. The application of such tenets differs in the above mentioned works but they nevertheless illuminate problems encountered by Glasse in his study of Huli descent ·notions. Glasse attempted to account for an “apparent contradiction in principles of descent” (1968:21) – a problem located at the level of cultural ideology – by reference to demographic features such as the “big proportion of male agnates and preponderance of non-agnates (1968:22) in the local community. The confusion of these levels of discourse – between cultural concepts and social patterns – is reflected in his use of phrases such as “local descent group” (Glasse 1959b:278) which has been shown by Keesing (1975) to merge what are analytically separable components. Later Melanesian ethnographers have uniformly urged that more attention be focused on the domain of cultural ideology and categories (cf.Keesing 1971:138, Strathern 1972:1) in relation to both statistical patterns and contexts of use as “rhetoric” Scheffler 1965:300). In this regard, the ascription of labels such as agnatic or cognatic is secondary to the concern of elucidating such principles and their function in social transactions.

The second perspective adopted here stems from the acceptance of ideology as “rhetoric” – a means of mediation between ideas and actions. Ideology is here defined as a set of ideas articulating a conception of social order, identified with the interests and policies of social groups, and inculcated through various educational institutions. A focus on the manner in which such ideas are verbalised points to the way in which linguistic disparities reflect fundamental social distinctions, inequities and imbalances. Where forms of social discrimination obtain – as between agnates and non-agnatic cognates or in male female
relations – these are likely to be marked linguistically by terms which embody statements about appropriate and inappropriate behaviour. This chapter treats language as a medium of discrimination focusing on choice of lexeme in different contexts, language about women as well as language used by women, uses of metaphor, euphemism and symbolic idioms, and the structure of category and group labels. The “language of conduct” is analysed to reveal a coherent set of beliefs about the nature and implications of female sexuality. The concepts and terms of genealogical relatedness and procreation provide an important point from which to begin the analysis. The salient kinship terms and idioms in Huli are as follows:
Bai leda – this is used to mark off cognatic from affinal relations (ref.D.10:speech by Daiya in pt.5 preceding: line 216). Bai means to be “tied” as in ho bai (to tie the struts of a house). The same semantic
element inheres in the term daba meaning 0 generation and which is invariably compounded with numerical adjectives in genealogical reckoning (ref.D.1:36-46). Daba denotes also a species of vine and, as in daba to the activity of binding/tying with cane.
Damene 1 expresses cognatic kinship and is most likely a compound of dama (ancestral spirits)+ne (generation marker: cf. note 2 Appendix 1). The inclusive aspects of the term are manifested in the word stem dama which connotes size as in iba dama (big rain) and damene (everything). Further specification of kin can be achieved by qualifying the above term with ainya (mother) or aba (father) to refer to maternal (ainya damene) and paternal (aba damene) kin respectively. The phrase “ainya damene” cannot be employed to denote sisters-sons as erroneously stated by Glasse (1968:26).

Tagini – this is used for cognatic kin with whom close contact is maintained (ref.D.9:419). The term occurs also in the construct wali tagini (incest) and is perhaps related to ta gira (outside) in the sense of bounding the range of prohibited sexual relations. Aria – in denotation it is co-extensive with damene as a general term connoting “relation”. The word is very probably a contraction of haria( track, path), a common idiom in the kinship language of Highland societies such as the Kuma (Reay 1959:65).

In addition to the above two further idioms commonly appear in disputes to signify notions of relatedness. The first is anda (house) which expresses the concept of incorporation as in andahane (house (anda) + inside (ha): a neighbour), anda hene ha ( he stays inside i.e. within the clan), anda ibu (com~ inside e.g. into the talk or conversation) and anda tambira (house (anda) + cold (tambe) – refers to a dead cognatic relative for whom a name prohibition obtains). The same concept finds expression in the phrase wi wandia (woman’s house) used to denote affinal relations (ref.D.4:1124-1125), and refers to the medial position of the female between her husband and her parents -‘one woman’s house’ symbolizing this relation. The second salient idiom is that of ‘bridge’ (togo) connection both in the sense of cognation (ref.D.4:862-863; 872) and again affinal ties (ref.D.4:1125). An understanding of these kinship expressions, particularly the concepts of “binding / tying”, is essential to interpretations of sexual intercourse and related spells explained later. As is the case with the Melpa term mei (Strathern 1972:8), the Huli word -ha-na can be used in a variety of contexts to signify a) child birth (honowini:carried(hono) + placed (wini);
b) transportation of live -animals in string bags (taba hana); c)payment of death compensation( agali hana: man(agali) + carried (hana); d) performance of bridal rites (hana:bag (nu)+ carried(hana): this is a reference to the black string bag ( mamai) which the bride uses to cover her breasts until after the birth of her first infant). A child is believed to be formed by the coalescence of maternal and paternal substances: a man’s semen travels down from the head in sexual intercourse to form the bones (kuni); a mother’s blood (darama) forms the flesh (dongone) and fat (habane). Following the child’s birth a man is considered contaminated for a period of about three months during which he is prohibited from giving food to children or from sight of his child. This state is signified by application of the construct agali gambaya. While the above formulation contradicts Glasse’s observation that a “child’s flesh originates from semen, its bones from menstrual blood” (1968:60), it is nevertheless consistent with a) concepts
of procreation prevalent in other Highland societies; and b ) the implicit meanings of the term kuni bone):

These connotations emphasise aspects of strength and permanency consistent with the kind of discriminations found in ideologies about male-female relations, and in references to agnatic descent as “one bone people” (kuni mbiyaore). This conceptual equation of maleness and bone figures prominently in D.10 where Daiya refers to himself as the “bone” (line 279) and to the death of his son as “my broken bone” (pt. 5 speech preceding line 216;330). 2 Discrepancies in the data presented here and that of Glasse, are related directly to misunderstandings about Huli concepts of procreation and language use. In an attempt to isolate the basic structural units of the society, Glasse cited the term hamigini (descendants of brothers 1968:23) – later changed to hamunigini (brothers-sons 1975:48) – which he understood to refer to both a cognatic descent category and to a local community which he termed a “parish ” . With reference solely to terminology, the correct Huli word is hameigini , a compound of hame (father) + igini (son cf. Appendix 1) i.e. a father-son unit paralleling Melpa structural categories tepam-kangemal (Strathern 1972:18). This failure to account for the agnatic bias in structural terminology may explain in part inconsistencies in earlier analyses which talked of agnation at higher and ambilineality at lower structural levels (Glasse 1959a), and later “the Huli are not organized into an elaborate segmental system” (Glasse 1962:167). A focus on terminology can be equally revealing when exploring the polysemic nature of the term hameigini in a comparative framework with the Choiseul lexeme sinangge (Scheffler 1965:40), and the Kwaio use of names like Kwafangi (Keesing 1971:125). These latter two terms denote variously a cognatic descent category, cognatic kin, task groups and a primary local and residential unit. The term hameigini in contrast refers unequivocably to a named patrilineal descent category, a semantic feature not shared by the word sinangge. This difference seems linguistically marked in the two cases: the compound hameigini (father -son) contrasting with the cognatic aspects inherent in the Choiseul word stem sina in sinangge (cognatic descent category), sinana and sinananggole (bilateral kin of +1+2+3 generations, Scheffler 1965:71). The following text illustrates the point I have made above:

Hameigini mbiyaore harali ogoni kuni mbiyaore ni howa igini damene honowinidago kuni mbiyaore gai bini. Hameigini hame mbiyaore laga.

The hameigini (clan) that is there is one, from the one bone all the sons are born, they are broken off from the one bone. It is said the hameigini is ‘one father’ (informant Hunguru: Koma clan). This phrasing is gain reminiscent of Melpa idioms which express the unity of descent in terms of the procreative functions of males. A range of other phrases expressing “oneness” can be gleaned from the various dispute texts. The clan is thus talked of as “one house”(anda mbiyaore -D.10: 125; D.7:310;169), “one garden, one land”(D.7:310), or a unit which shares food as expressed by the idiom “one piece of pig 0 (D.7:169). Where the word “penis”(wi) occurs in speech it tends to reflect interpersonal elements of kinship as in D.11:120 and D.10 where Daiya talks of his son as his “penis” (253), “wi irane kogo -the stem of the penis is me”(278). The same idiom can be used to relate or incorporate non-agnatic cognates (ref.D.10:49-52), as well as
discriminate against them: “yamuwini ti wi tarame honowini non-agnatic cognates are born of a different
penis” (field notes). Techniques of incorporation and alienation as used in disputes are discussed more fully later, and it is sufficient here to note the dual way in which such idioms can be used in rhetoric. The implications of “penis” as a procreative symbol are inherent in the phrase “wi ti” (straight penis) signifying a confused state of mind as when sexually excited (ref.D.10:407-408), and as a synonym for ‘trouble’ (ref.D.10:382-383;D.7:41,50).

Female agnates are referred to collectively ashame wane (fathers-daughters) and male agnates as aba-igini (father-sons). The following text encapsulates the important relations which are conceived between paternal substance, patrilineal descent and land. It illustrates the way in which notions of descent and locality are fused in single terms:

Ira tene harua dagua ina hameigini timbuni kemagoni irane magi manda beregoni. Ira ginigene udu kadagua o ina hameigini emene haradagua. 0 mani honowima bibai lama bedago ira gene udugo ale o erebabi holebira habi bulebira ai ira tene timbuni ereba nahole. Ibu ira mbira henegoria ibu delaabo halu haabo haradgua holene ira tene ogoria haradagua holene nde agali timbuni maru kagoria; o dindi tene henegoria kemagoni.

The clan (hameigini timbuni) stays like the base (tene) of a tree, we are staying together like the trunk. The sub-clans (hameigini emene) are there like the branches (ginigene), those who are born afterwards continue to stay like branches. They may get lost or they may stay; the main base (tene) will not get lost. The first tree that is there will continue to bear fruit (delaabo), the base stays like that, as the first man on the land. We are staying on the ancestral land( dindi tene). (Informant Gurubugu:Koma clan).

The segmentary structure of the clan (hameigini) is equated metaphorically with the trunk or base (tene) of a tree and its branches, but whereas in Choiseul the same metaphorical distinction is utilised to mark segments of a cognatic descent category (Scheffler 1965:45) in Huli the model is an agnatic one. The same order of divergence characterizes the conceptualisation of genealogical charters. In Choiseul this knowledge is called tutusi sinangge (Scheffler 1965:62) and genealogical structure is clearly cognatic; the Huli malu, by contrast, is always patrilineally structured (ref. D.1). This is again linguistically reflected in the term malu itself as compounded of ma (as in mama – +2 generation male cognate)+ lu (long, i.e. a list of ancestors; cf.D.1.). Ideally only the headman of each clan – an agnatic descendant holds the complete version of this knowledge, and may use the appropriate speech genre in disputes. The occurrence of the words timbuni (big) and emene small) to discriminate structural levels, parallels Melpa usage where such terms can operate “right down the segmentary scale, higher levels being bigger than lower ones” (Strathern 1972:43). The term tene, which can function as a category label for all “agnates”, may be considered to fuse in the above text notions of patrilineal descent and residence – agnates are source (tene) as well as residing on the land like the base (tene) of a tree. Polysemy typifies most other idioms relating to land rights. Examination of Table 1 reveals the consistent male orientation of phrases used to refer to concepts of procreation, descent and residence. Each term – aba, kuni and tene – is capable of expressing notions of kinship and locality, and hameigini itself signifying both a patrilineal descent category and the associated territory and group thereon residing. An agnatic model thus provides the org anizational structure of the parish whose segments are named after putative sons of agnatic Ancestors; only agnates are entitled to use the patronymic suffix that constitutes the relevant clan appellation.
By comparison, the implications of “maternal blood” seem far more complex in Huli society. Those cognatic kin who trace relationship to agnates through a female tie are termed yamuwini. The full import of this usage is developed later but at this juncture we may note the following common formulation:
Yamuwini ti aba tara, ti hameigini tara
Yamuwini, their father is different, their clan is different.

Such statements are intelligible only by reference to the dual nature of the term hameigini elucidated above. Yamuwini are thus members of the local parish community (hameigini) but not necessarily the agnatic descent unit after which the territorial formation is named. They are not ‘one father’ with tene agnates). The idiom of “one blood”(darama mbiyaore) has a number of important similarities and diff-erences with the corresponding Melpa construct mema tenda (Strathern 1972:15). Not only do the Huli metaphorically refer to sub-clans as “branches”, but there is also a recognition that female agnates’ blood creates “branches” as in Melpa. This is implicit in the kinship term magane (mothers-brother) which is also a generic word for “tree branch”. In similar fashion to its Melpa counterpart, the Huli idiom of one-blood can signify both agnatic descent (ref.D.2:33), and its inclusive interpersonal relationships (ref.D.2:15-16) ,in addition to a range of cognatic kin. Despite these similarities, however, the extensions of the phrase are quite different; in Huli cross-cousins’ children can be said to be of “different blood” (darama tara). In order to facilitate understanding of this dogma I intend to interrelate certain aspects of marriage norms, kin terminology and social structure with a focus on the “interpersonal aspect of genealogy” which La Fontaine (1973: 44) found wanting in New Guinea ethnography. The complex of relationships and associated terminology with which I am concerned are set out in Figure I. and are central to any comprehension of Huli kinship and marriage behaviour.

to be continued…

(An extract from Talk Never Dies: An Analysis of Disputes Among the Huli. A thesis submitted by Laurence R. Goldman for the degree of Doctor of Philosophy, University College, London. February, 1982., p. 21-35.)

  1. This may be related to the Duna damene’ (Modjeska 1977:7), though in Huli it expresses cognatic kinship not cognatic descent. []
  2. Reference may also be made to D.11:106-107. []

Tagged in: