We can see now that the seven-clan system that exists in Porgera today is only tenuously related to the sociality that existed prior to the arrival of mining in the valley. The question then becomes how the entrepreneurial agency of Ipili network building was transformed into the segmentary descent groups of the seven-clan system. How, in other words, did the mine’s interest in the Ipili elicit this transformation on their part? A full answer to this question would have to take into account the long history of medium- and small-scale gold mining in Porgera which served as a crucial backdrop for the creation of large-scale mining in the 1980s. Here I will focus on one key moment in this process — the period from late 1985 to 1989. This period begins with the official decision to create an authoritative list of ‘who the Porgera gold mine landowners are’ and ends with the signature, on their behalf, of the Porgera Agreements by their 23 authorised agents.

In 1983, exploratory work at Porgera uncovered a zone of particularly rich ore, and the following year an ultra high-grade area within this zone was discovered. This work, as well as developments in metallurgy and financing, made the creation of a large-scale gold mine in Porgera seem increasingly likely (Jackson and Banks 2002: 119–38). Much was still unknown about how, concretely, the mine would be created, financed, and regulated. The late 1980s were spent attempting to clarify how these issues were to be dealt with. One of these clarificatory projects was the Porgera Land Study.

In order for a mining lease to be issued, the owners of the land in question had to be identified and compensated. To this end, on 26 September 1985, the Secretary of the Department of Enga issued an order for public servants to begin a land study to generate a list of customary landholders. The PNG Department of Minerals and Energy created a position of ‘Porgera Coordinator’ with responsibility for liaising with all relevant ministries regarding the mine. John Reid, a former kiap (government official), was appointed to this position. He in turn oversaw a group of Papua New Guinean kiaps who conducted censuses of the area. Their guideline for conducting the land study was a thin photocopied booklet produced by a senior public servant in the national department. This specified that they were to record the names of adult men as ‘landowners’, and that each of these men could be associated with one and only one clan.

By December 1986, the land study had collected 15,000 names of Ipili in their schedule of ownership. This was an astonishing feat considering the fact that, by their own estimate, there were only 3,000 Ipili living in the valley at this time. The land team was flummoxed: areas of two or three hectares were being claimed as the customary land of more people than could ever live on the land or, in the case of very small plots, even physically stand there. It appears that individual Ipili were responding to questions by being maximally inclusive and listing their entire kindred.

Both Ipili and government representatives were unhappy with this state of affairs. At a meeting held in December 1986, government officers complained that people were registering themselves multiple times on multiple plots of land, and suggested that Ipili ‘select one or two members of a family to represent them in other Landownerships [sic] in the other clans/sub-clans’.^[10] For instance, even if a man was married to a woman in another group, his claims to that group’s land (through his affinal status) would be represented by his father-in-law or wife. This proved completely unacceptable to Ipili, who saw this as a radical reduction in their portfolio of relationships. As one man put it, ‘Porgera people have Landownership Rights in more than one clan. Therefore we want all our names to be enrolled or registered in all the clans that we own Lands’. Another ominously noted that ‘if any names are excluded, there will be troubles’.

Eventually a proposal was put to the meeting by Kurubu Ipara, a Porgeran who had previously worked as a kiap and was at that time working for the mine’s exploration team. He suggested a solution that would ‘make it easy on the kiaps’. Instead of proceeding by visiting a piece of land and eliciting the names of all the people associated with it, the land study would instead proceed on a ‘clan by clan’ basis. The result would be a series of seven ‘schedules of owners’ that would be attached to the land study. This would allow the government to retain the idea that there was a set of distinct ‘clans’ which owned land. However, each individual Ipili would be allowed to list themselves in as many of these ‘clans’ as they considered themselves to be part of. This would allow Ipili to retain the inclusivity that they desired. Indeed, in some ways it was more inclusive than previous practice, since Ipili could now be fully ‘in’ clans to which they had previously only a potential and possibly tenuous claim.

John Reid, who was at the meeting, objected that a count done in this way no longer tracked the actual population of the Special Mining Lease and would result in the sort of thing that the meeting has originally been established to avoid: a list of 10,000 people that described an actual population of 3,000. Ipili assured him that redundancies would not over-inflate the list because ‘in the previous investigation we have included all our wantoks (people who speak the same language) who are living outside Porgera and therefore the population increased. Now we are cutting it down or restricted to landowners living within the Porgera valley’. The Ipili present at the meeting suggested that, in exchange for the government’s recognising multiple ‘clan’ affiliation, they would limit claims to membership — out-marrying women, their husbands, and their children could be included on the list, but all other affines would not be entitled to membership in landowning ‘clans’. As a result of this agreement the cognatic stocks which had previously been an important part of Ipili sociality now became ‘clans’ and these became the sole form of sociality recognised as appropriate for true ‘landownerhood’.

As a canonical account, the social organisation of the seven landowning clans composed of 23 sub-clans has shaped life in Porgera for roughly two decades. It exists as a taken-for-granted fact about the valley. The original land study is not only rarely consulted; it is quite difficult to find. Close examination of the original document, however, reveals the traces of the more unruly arena of entrepreneurial agency which, on paper, it replaced.

For example, in the case of the Waiwa ‘clan’, the schedule of owners lists two sub-clans — Waiwa Yaliape and Waiwa Lunda (GoPNG 1987c). This in itself is not a surprise, as the Lunda are a large group who are associated with a piece of ground known as Upalika, whose members include several prominent Porgerans and the wife of one of the key coordinators of the land study. The Lunda are not, however, the ‘owners’ of any land within the Special Mining Lease; Upalika lies outside it. It is not clear, then, why they ought to be included on a list of owners or have an agent who signed the Porgera Agreements of 1989, as the agents for Waiwa Lunda did. Furthermore, it is not clear who actually ought to be agent for Waiwa Lunda — the agent for Waiwa Lunda listed in the land study is William Gaupe, but the agent who signed the Porgera agreements is listed as Pospi Karapis (Derkley 1989). Did Gaupe delegate his authority to Pospi? There is certainly no record of that fact, and during the time that I knew him William certainly did not make any such claim.

A similar situation pertains with regards to the Tuanda. The Tuanda are divided into two ‘sub-clans’, Ulupa and Yapala, which are represented by two agents, Sole Taro and Ambi Kipu (GoPNG 1987b). One of the Porgera Agreements, however, lists Aiyope Yawane as the signatory for Yapala rather than Sole Taro, and to confuse matters even more, the ‘signature’ on the document is a thumbprint with the word ‘Sole T.’ written over it (Derkley 1989). This is unusual, since Sole is proud of his education and is one of few agents who could sign his own name.^[11] In the agency delegation document authorising Sole as agent for the Tuanda (as well as in other documents), he uses a florid and elaborate signature (GoPNG 1987b).

Furthermore, there is no coherent account of the segmentation of the Tuanda. The terms ‘Ulupa’ and ‘Yapala’ do not appear in the ‘sub-clan’ entry on the top of each page of the schedule of owners of Tuanda. Instead we have ‘Kareya’ and then a series of what are presumed to be sub-subclans in parenthesis — Kareya (Aiyengi), Kareya (Amini) and so forth — as well as a handful of other names. Most tellingly, the sub-clan ‘Marinaka (Lio)’ is included, and its ‘address’ is listed as ‘c/o Catholic Mission Kasap, Yangiyangi Village, Mulitaka Patrol Post — Lagaip District’. ‘The Marinaka’ are in fact part of the larger clan diaspora of which the Tuanda are also a part. They come from Mulitaka, an area east of Porgera which is today considered ethnically ‘Engan’ rather than ‘Ipili’. The schedule lists 94 Marinakas: 87 in Laiagam (the township near Mulitaka), and seven people who live in Porgera proper, including one Marinaka man, his wife, and three children. In other words, the schedule of owners purports to list a segmented series of sub-clans, but embedded within it is a specific form of regional sociality — a small group of Engans from Marinaka had moved to Apalaka on the basis of their diasporic ties to Tuanda, recorded themselves as landowners, and then enumerated their kin in Mulitaka as landowners too. Given the agentive nature of kinship in this region, it is not surprising that these people are included in the schedule of owners — as we have seen, the Waiwa Lunda are included on the schedule of owners as well as in the land study itself, despite the fact that (like the Marinaka) they have no land within the Special Mining Lease. But given the imperatives of the system of agency, it is no surprise that no Marinaka agent was appointed and that the Marinaka do not appear as one of the sub-clans in the land study or the Porgera Agreements. Between the compilation of the schedule of owners and the completion of the land study they had lost their status as landowners — a point that the Marinakans living in Porgera during my fieldwork have not forgotten.

Examples of this sort could be given in many of the other clans.^[12] In closing I will consider only one of these — the land study’s attempted segmentation of the Pulumaini. Some individuals living on Pulumaini lands did not feel comfortable with the idea of agents at the ‘sub-clan’ level and instead opted for a more granular level of representation, but others made a different choice. The result was a mix of different-sized groups and delegations of agency which simply could not be subsumed under a lineage model. In the agreement between the National Government and the Porgera Landowners, there are six Pulumaini ‘clans’ — Ambo-Wagia, Ambo-Amu, Ambo-Endewe, Ambo-Gai, Ambo-Paramba, and Ambo-Yuga (Derkley 1989). Based on these names, each of these groups appears to be a subset of a larger ‘Ambo’ group. In the agreement between the Porgera Landowners and Enga Province, there are only five sub-clans, none of which are subordinate to Ambo. The clan groups Amu, Endewe, Gai, Paramba, and Yuga are listed, and one of these, Amu, has been added at the bottom of the document in pen (Derkley 1989). Neither of these two agreements match the land study itself, which lists the Pulumaini as being composed of Ambo-Wagia, Ambo-Amu, Tokome, ‘Pulumaini Sub-clan Yamawe’ (with the word ‘Ambo’ pencilled in between ‘Pulumaini’ and ‘Sub-clan’), Ambo-Napali, Ambo-Endeme, Ambo-Gai, Yamili-Wapini, ‘Pulumaini Sub-clan — Paramba’ (with the word ‘Yamili’ written in pencil between ‘Pulumaini’ and ‘Sub-clan’), Pariwana, and Yunga (GoPNG 1987a). In other words, these documents have never presented a coherent model of the Pulumaini as a clan.