Matrilineal Kinship at Sea in Bougainville, Papua New Guinea

This paper explores matrilineal kinship in the Buka area, in the Autonomous Region of Bougainville, from the perspective of saltwater people on Pororan Island. In Bougainville and elsewhere in Melanesia, anthropological research has highlighted the importance of joint work in the gardens, of sharing and exchanging garden food, and of negotiations of access to land for kinship and relatedness in the region. Where does this leave saltwater people, who often have only small areas of land of their own, take little interest in gardening and depend on traded sweet potatoes or imported rice for meeting their subsistence needs? In the first part of this paper, I indicate “landed” bias in anthropological research on kinship, including matrilineal kinship. I then suggest complementary descriptive and analytic terms that may be useful for researchers who want to understand kin relations among saltwater people, based on my experiences among Pororan Islanders in Bougainville. Finally, I indicate the theoretical contribution that these terms can make to research on kinship in landed settings, as well.


INTRODUCTION
The topic of this paper is matrilineal kinship among Pororan Islanders in Bougainville, PNG, and its distinctive 'saltwater' quality. During ethnographic fieldwork, mostly with the Pororans but also with some of their land-based relatives and neighbours, I found Pororan Islanders' ways of talking about and engaging in kin relations to be significantly different from those of their land-based kin, and also different from what I expected, based on my reading of the Melanesian ethnographic literature. I suggest that this difference is linked to their particularity as self-declared 'saltwater people'. Studying kinship among saltwater people requires a different analytic vocabulary from the developed through research with landward-oriented people. At the same time, what we can learn about kinship from saltwater people can enrich our understanding of kinship "on land", as well.
The main part of the paper is based on observations from ethnographic research in northern Bougainville in an area locally called Buka in 2004-05. The Buka area includes Buka Island, locally called the mainland, and a number of much smaller islands. 1) Pororan is one of them. The original aim of my research was to find out how saltwater people in Buka would be affected by changing notions of land in Bougainville just after the Bougainville Crisis, a protracted conflict triggered by landowner dissatisfaction with the Panguna Copper mine in central Bougainville. 2) The Crisis ended officially in 2001, and my field research fell into the period of a formal, UN-monitored peace process that lasted until mid-2005. 3) As a result of the Crisis, Buka people were aware of land being a source of riches, but also of conflict. Many of the peace-building activities in the villages were explicitly designed to address this danger. The central question of my ethnographic research with Pororan Islanders was how this heightened awareness of land as a (potential) resource and source of conflict would affect people who traditionally owned hardly any land at all, and who depended on access to the land of others for their livelihood.
Official statistics indicate that there were about 2500 saltwater people in Buka, out of a total population of 35,000 in Buka and 175,000 in Bougainville. They lived on four small islands just west of Buka Island. Especially on the northern two, Pororan and (much smaller) Hitou, soils are very poor and population densities high (above 600 per square kilometre). The inhabitants of those islands cannot subsist on the cassava and sweet potatoes they grow in their small gardens on Pororan (Bourke and Betitis 2003). Traditionally, the islanders have bartered fish for starch food with relatives on the Buka mainland. 4) This "traditional barter system" continues, and direct barter is complemented with cash transactions. 5) In 2004-05, the Pororans acquired most of their starch food at these traditional markets held regularly on the Buka west coast. Up until mid-2004, many of them also made gardens on land belonging to their matrilineal relatives on Buka, as well as on the land of a local mission station. However, these arrangements were disrupted during my fieldwork by disputes among the mainlanders over this land. Finally, many Pororan Islanders, mostly men, were engaged in wage labour Buka Town during my fieldwork. They worked as carpenters or for one of the hotels that catered to observers of the peace process. They came back every other weekend and brought rice and tinned fish. When exchanging fish for starch food with Buka mainlanders and when negotiating access to land on the Buka mainland, the Pororans say that they 'trace back the footsteps of the ancestors'. They retrace physical movements and relations that date back to the times of ancestral migration. Buka people are said to have come from a particular mountain on Buka, called Punein. After an argument, different matrilineal groups made their way down this mountain and around the Buka area, each along their particular road. Eight or nine of these matrilineal groups were represented on Pororan during my fieldwork (depending on how one counts). Their members are the descendants of people who were sent to the island to catch fish for the rest of their group, who were based on the mainland. It is the responsibility of the tsunon and hahini, the male and female traditional leader in the matrilineal group to keep in touch with their counterparts on the mainland. They meet with them occasionally for telling migration histories and exchanging genealogies that can clarify how exactly the people at their respective places are related. Updating knowledge of relations regularly is important, say the Pororans, for finding 'water, food and help on the road' when they travel on the mainland. It is also important, they say, to be clear about who the people are whom one gives fish to at the markets, and whom one asks for access to garden land. Obligations attributed to (ancestral) kin relations are often invoked in these transactions and negotiations.
The hamlets on Pororan are no longer fishing camps, as they were in ancestral times. They are permanent settlements, and there are even garden areas on the island, in which each matrilineal group holds a section. Still, the impression of a 'saltwater place' remains, and visitors often point this out. Many hamlets on Pororan are located on the beach, and none are no further than 5 minutes by footpath away from the sea. Pororan Islanders' daily rounds and activities, too, are sea-oriented. With the exception of the very smallest, everyone baths in the sea twice a day; children and adults, men and women spend a lot of time gathering shells on the reef or fishing in the lagoon from canoes or using goggles and a spear; and they pride themselves in being great fishers and lazy gardeners. Finally, in addition to migration stories that recount the different matrilineal groups' descent from the mountain of Punein, the Pororans tell another origin story, as well. According to this story, the first human beings were a sibling pair walking around on the reef. The question that motivates this paper is what difference their saltwater place, sea-oriented activities and myths make for the ways in which kin relations are enacted and conceptualized. This seems important for two reasons. First, for me as for many other ethnographers, learning kinship was the sine qua non for learning anything at all in the field. However, a 'landed' bias in the ethnographic literature on kinship in Melanesia has made the process of learning kinship from saltwater people a rough ride. Second, toward the end of the paper, I will suggest how ethnographic observations of saltwater kinship could add to our understanding of kinship in "landed" settings in Papua New Guinea, as well.
To take my main argument ahead, Pororan Islanders' ways of apprehending, engaging in and negotiating kin relations thrive on attention to the physical movements of people on and around the island. The islanders learn about kin relations by observing the movements of those around them, by asking passers-by where they are going and discussing observations and answers with one another. They engage in kin relations by carefully adjusting their own movements so as to make specific relations visible, and by interfering in the movements of others. I suggest that theirs is a distinctively saltwater mode of learning and doing kinship, congruent with the perceptive habits of people trained to a significant portion at sea, and in surroundings in which most things are in motion, for most of the time. By contrast, much of the literature on kinship from Melanesia details modes of doing kinship grounded in the perceptive habits of people who live and work on land. Among these land-dwellers, sharing substance makes people 'the same', and the separation of substantive flows or the transformation of substance makes them 'different' (Wagner 1967). Theirs are gardeners' modes of apprehending and learning kinship (Leach 2003). "Landed" models have a powerful presence in the ethnographic literature on Melanesia, but offer little guidance for understanding kinship among saltwater people.
The paper is divided into three parts. In the first ,   I highlight the importance of substance for kin relations  among land-oriented people and show, with reference  to some ethnographic examples, how these substances  relate people eventually to the land as well as to each other. On this background, it becomes necessary to ask how -and if -kinship among saltwater people is "grounded" in its surroundings, and what analytic approaches to kinship such a saltwater "grounding" calls for. This question is addressed in the second and main part of the paper. Here I provide details on Pororan Islanders' ways of perceiving kin relations in the movements of people, on the methodological challenges this raised for me during fieldwork, and on the specifics of the connections through movement that people maintain not just with one another, but also with their surroundings. In the third and final part, I argue that the contrast between landed and saltwater kinship thus set up is not absolute. Taking inspiration from some Pororan Islanders' ability to switch, on certain occasions, from apprehending kinship by tracing human movements to discussing kinship in terms of genealogical relations and relations to land, I re-visit some analyses of kinship among land dwellers from a saltwater perspective.

LANDED KINSHIP: THE POWER OF SUBSTANCE
Substance has become a central term in the literature on kinship in Melanesia since the 1970s, and in the context of broader disciplinary debates about how to investigate kinship without becoming trapped in Euro-American folk models (see esp . Schneider 1968). In analyses of kinship in Melanesia, the focus on substance emerged as an alternative to "African models" of descent in the Papua New Guinea (PNG) highlands (Barnes 1962). Ethnographers working in the recently opened PNG highlands in the 1950s and 1960s reported a dogma of descent that seemed to match the African models. However, they also found shallow genealogies (Salisbury 1956) and flexible structures (Brown 1962) that put the salience of the model into question (see also Langness 1964) Theoretical responses to this and similar observations differed. Langness' (1964) conclusion from his observation among the Bena Bena was that "the sheer fact of residence […] can and does determine kinship" (ibid.:172). This resonates with later arguments for opening the field of kinship up to include various modes of 'relatedness' (Carsten 2000). 6) More influential at the time, however, were two other ethnographers' responses to this and related ethnographic problems. One of them, Roy Wagner (1967) inverted a core argument of Levi-Strauss' (1969) and argued for the Daribi that descent (or consanguinity) related people to one another, but that group definition was achieved through exchange. The other, Andrew Strathern (1973) remained close to the original formulation of a problem between ideology and structure (or transactional patterns). Drawing on earlier suggestion by Scheffler and Keesing, he asked what would "bring ideology in closer alignment with transactional patterns" (A. Strathern 1973:26). In his Melpa ethnography and in other ethnographers' work (Salisbury 1965;Wagner 1967), he found an answer: substance. With respect to the Melpa, he suggested "[…] clansmen share substance in some way through their descent from an ancestor. Another way in which they share substance is through the consumption of food grown on clan land" (A. Strathern 1973:29).
The argument that both descent and sharing food from the land are ways of sharing paternal substance has been made productive since then in analyses of both kinship and gender. With respect to the latter, the different fate of paternal and maternal, male and female substance has been of special interest. Paternal substance, passed down the generations in procreation and augmented through the consumption of food grown on clan land, constitutes patrilineal kinship. Maternal substance, though transmitted in procreation, as well, is later eliminated from children's bodies in various ways. Among some groups, it is "returned" to the mother's relatives in the form of marriage and child payments (e.g. Wagner 1967). These exchanges also create the differences between groups linked by genealogical relations that, with the advent of structuralism, could no longer be taken for granted (Wagner 1967(Wagner , 1977. In other cases, boys undergo special procedures during male initiation for eliminating female substance from their bodies and replacing it with male substance (e.g. Godelier 1986;Herdt 1981;Herdt and Poole 1982;Kelly 1993;Salisbury 1965). After initiation, female substance is polluting to them (e.g. Meigs 1976;. More recently, particular understandings of gendered substance, its circulation and its healing or harmful effects have been pursued further, among others in studies of sexually transmitted diseases (e.g. Wardlow 2002;Wood and Dundon 2014).
Besides the gendering of substances, from the 1990s onward, many ethnographers have become increasingly interested in the role that land plays in the processes that generate kinship in New Guinea.
Land is an integral component in the processes of sharing substance (e.g. LiPuma 1988; Merlan and Rumsey 1991), and places are equivalent to bodies as repositories and conduits of this substance. It is not "a kind of container in (or on) which life goes on", but "it enters directly into the constitution (generation) of persons" (Leach 2003:30). It follows that not only interactions among people are constitutive of kinship, but also the engagements of people directly with the land. Not only feeding (e.g. Battaglia 1985;Munn 1986;van Heekeren 2004;von Poser 2013) but also gardening needs to be studied if one wants to understand kinship (see esp. Leach 2003:91-125).
Writings on both gender and "growth on the land" testify to very particular understandings and processes at work in different ethnographic cases, not all of which can be illuminated with reference to the term substance (see also M. Strathern 1999:49, 264n5). Sandra Bamford (e.g. 2004) has argued that substance has been over-rated in studies of kinship in New Guinea. Her Kamea hosts, belonging to a language group famous in the literature for a strong concern with bodily fluids, do not conceive of kinship as a relation of substance, neither between parents and children, not between siblings. While siblings are said to be "one blood", the term is a reference to the "shared experience of having been nurtured within the same woman's womb" (ibid.:291). While inter-generational continuity is indeed achieved through work on, and knowledge of the group's land, there is no indication that it depends on a substantial connection between people and their land. Thus, while substance remains central in accounts of kinship in New Guinea (and beyond), there are indications that kinship is not necessarily a relation of shared substance.
In the following section, I present ethnographic material from Pororan Island that shows that these saltwater people are not interested in detecting flows of substance, either among people or between people and their land. Rather than flows of substance, they carefully observe the physical movements of human beings that constitute kinship, and that allow others to "see who is who" on the island, as the Pororans say.
I attribute their interests in human movements rather than substance to the particular perceptual habits of people who live in surroundings in which places are less readily perceivable as conduits of substance than as assemblages of human and non-human elements that float along, get stuck together for a moment, come apart and move on again.

SALTWATER SURROUNDINGS
Pororan is a tiny place packed with people, "what if it will sink", as the islanders sometimes say.
Population statistics give an impression of the density of settlement: there are about 1,225 people on about 2 square meters. However, these people were far from stationary during my time in the field. Eight youths were away for high school in Buka Town. About 20 people, almost all men, were working in Port Moresby or in other urban centres in Papua New Guinea. Some of them returned once a year, but others had not been back for several years. Conversely, a handful of men from East New Britain Province had arrived on the island several years earlier, and while some of them said they were just in transition, others had married on Pororan, had children, and showed no signs of leaving again. Besides these more long-term, longdistance moves, about 40 Pororan Islanders, again mostly men, were working in Buka Town during my field research and only came back every other weekend, weather permitting. In addition, several women were staying with mainland relatives for four months, waiting for a sweet potato garden they had made on the mainland to be ready for harvest. Moreover, there was a less predictable coming and going of adults and children alike who visited relatives on the mainland, for as little as a day or as long as a month or two. Sometimes they went 'just for fun' or 'for a little holiday'; more often, however, they went to request 'help', that is, starch food or permission to make a garden on land belonging to mainland relatives and stay at their hamlet till it was ready for harvest. Finally, Pororan Islanders enjoyed 'going around in Town', that is, Buka Town, for shopping or just strolling around. Sometimes they combined trips to Town with visits to mainland relatives. The same held for the traditional markets on the Buka west coast. Usually, there would be several islanders who, often spontaneously, decided to join their mainland relatives for a couple of days after the market. They would return via Town or on the next market day. Thus, there was a constant back and forth by dinghy to and from the island. Eight dinghies owned by Pororan Islanders were active at the time of my fieldwork.
The islanders could tell them apart by the sound of their engines.
In their frequent 'going around', as they called it -leaving the island for 'finding food', for work or 'just for fun', and returning, in unpredictable patterns -the Pororans were continuing the habits of the ancestors, they said, who were going back and forth between their fishing camps on Pororan and their relatives' place on the Buka mainland. The Bougainville Crisis had temporarily disrupted these habits. However, by 2004, the Pororans were busy again 'going around in the bush' (visiting mainland relatives) and 'going around at sea' (fishing). They were also going around a lot on the island itself: moving in with relatives at another hamlet for a couple of days or weeks in order to look after kin who needed help, because relations at their own hamlet had become a little tense, or for no reason in particular (or none they wished to share with me). This was 'island life', they told me: not knowing when one would be where, if one would 'find money' or not, if one would be able to obtain sweet potatoes for one's fish at the market, when someone supposed to bring supplies to the island would arrive, or if someone whom one wanted to bring a basket of sweet potatoes would be at home.
When I asked Tsireh, the tsunon, the traditional leader who was responsible for me during my stay on the island, if he didn't mind people coming and going in manners that made it very difficult to plan any cooperative undertaking, he said: 'That's how it is over here, at sea'. He pointed to a piece of driftwood. 'You see', he said, 'this one, too, comes and goes. Who knows where it came from? It got stuck here three days ago, that's when I first saw it. Now it is stuck here. But who knows what will happen next? Maybe tomorrow, the tide will carry it away again. Tsireh drew for me between 'our ways here at sea', human movements, the movements of driftwood, and the processes of island growth and diminishment were complemented later on by commentaries about a basic likeness between humans and fish. Fish 'go around at sea' in order to 'find food', just as the islanders do when they go out in their canoes in order to catch fish. The movements of people and the movements of fish are equally unpredictable, according to the Pororans. Pororan fishers, including highly and consistently successful ones, insisted that they "just tried", but never knew in advance if the fish would be around at the spot at which they were looking for them or not. When talking about fishing, the islanders emphasized their sense of surprise at 'bumping into fish', that is, feeling the weight of a fish on their hook, rather than their satisfaction at achieving the aim of catching a fish.

Movements of fish were only of interest to
the Pororans while they were at sea to catch it; by contrast, they were interested in the movements of other persons any time. They could spend hours and entire days observing and discussing the comings and goings around them: who had gone where, with whom, for what stated or likely purpose. From these observations and discussions, they drew tentative conclusions about social relations. Important among these were relations of kinship, especially relations of matrilineal kinship.

APPREHENDING KINSHIP IN MOVEMENTS
So how did the Pororans apprehend kinship in the physical movements of people around them? There are several aspects to this, and several strategies I learned from them. First and most basically, Pororan Islanders employed their eyes and ears for registering who was going where, from where, with whom and (most likely) to see whom. I learned paying attention to movements in this way from women sitting in front of their houses in the mid-day heat, chewing betel nuts and apparently doing nothing for hours on end. They watched people as they went about daily chores, called long drawn out 'ooo's out to passers-by, and listened to the 'ooo's of others that made it possible to track a person's movements beyond one's field of vision.
Second, the Pororans used and carefully differentiated between many different verbs of movement when discussing everyday affairs, when telling stories of the past, or when commenting on the current state of social relations on the island. For instance, they differentiated between someone having 'gone to Town' ( ela niTaun, Hapororan) for a particular, known purpose) or 'going around in Town' (eroror iTaun, Hapororan) for no such purpose). Furthermore, married women on Pororan 'returned' from their husband's halet, where they lived, to the hamlet belonging to their matrilineal group, although they had never lived there; however, they 'went to' the hamlet of their paternal relatives, at which they had grown up.
Finally, the Pororans used gesturesmovements of their arms and hands or eyes and heads -to highlight movements of particular shapes and qualities. Once I had learned to identify those, I could detect relational meaning that people attributed to others' movements even in the absence of any verbal commentary, just by looking at the movements of their hands, arms or heads. As I myself learned to employ these linguistic and gestural forms (more or less) correctly, the Pororans began to draw me into more sophisticated, and sometimes more esoteric discussions of movements past and present.
The movement that characterizes relations of matrilineal kinship, both relations between women and their children and the relations of people to their matrilineal group, is a smooth downward-outward and inward-upward arm movement. Here is Tsireh's formal account of matrilineal kinship, bihainim mama [following one's mother, Tok Pisin]: 7) Matrilineal kinship is just that: following your mother. In the beginning, the child goes around during the day, and then comes back to the mother every night [arm movement downwardoutward and inward-upward]. When they grow up, children leave their mothers [arm movement downward-outward]. A girl will go to stay with her husband [arm movement downward-outward]. She will stay at his place, and her mother-in-law will look after her. At first, she will just stay, she will not come back. Once she has settled in, she can come back and visit her own people. She visits her mother, who is living at her father's place, and all her relatives there [arm movement inwardupward]. She will bring food for them, or a little something from her husband's place. When she has children, she will bring her children, to show them to her relatives [arm movement inward-upward ]. Now, when her father dies, her mother will return to her own place, her mother's place, the place of her pinaposa [matrilineal group, Hapororan; arm movement inward-upward, higher up and in towards Tsireh's face than the first]. The daughter will then go to visit her there, to support her, maybe by bringing sweet potatoes or making firewood for the mother. This is good; this is showing respect to the mother and to the ways of our ancestors. When the daughter grows old and her husband dies, she will follow her mother [arm movement inward-upward] to her mother's place, along with all the other women of the same pinaposa. Pinaposa -their children are stuck together, that is the meaning. That is it. That is matrilineal kinship.

(Interview Cyril Tsireh, September 2004)
Pororan mothers use the same arm movements and the same verbs of movements that Tsireh used when he gave me this explanation when they interact with their babies and children. During my first week at the hamlet of Lulutsi, I watched Tarasih, the smallest grandchild of the hamlet's senior woman Salu, as she was being taught to walk. Tarasih was encouraged to 'go now, go to grandmother', 'go now, go to your brother', but most importantly, after every excursion to another person, to 'come back now, come back to mother'. Tarasih's older siblings, two boys aged nine and four, 8) meanwhile, were free to leave after breakfast and 'go around' the village, as well as to go around the lagoon in their little canoes. However, their mother admonished them to always 'return to your mother' for meals and for the night. In the evening, her children sat with her and their grandmother around the fire. The women would ask what they had been doing, where they had gone, and would encourage them to play with some people's children but not others, on the grounds of particular histories of mutual help with their relatives. Besides returning to their mother at certain intervals during the day and at night, Pororan children, and especially girls, are also expected to accompany their mother when she returns to her mother, or to the place of the mother's mother's matrilineal group. Jocelyn, Salu's daughter-in-law who was usually staying at husband's place, regularly left the island for a couple of days to 'return' to her mother's hamlet on the Buka mainland. Usually, she would take her daughter Hessa along, 'so that she will know where her place is, where her mother's people are'. By taking her along on her trips to the mainland, Jocelyn was effectively making Hessa a member of her matrilineal group: she was making her display the movements that made her recognizable "a woman from Nova", her mother's village, and within it, a woman from the hamlet belonging to her mother's matrilineal group, even as she was growing up with her father. Jocelyn also took Hessa along when she visited her father's place, at another village on the Buka mainland, where Jocelyn had grown up. She hoped that Hessa would find a husband in that village, where Jocelyn's paternal relatives would support her. These visits, however, were visits and not 'returns'. They were less important for Hessa than returning to her mother's people, as Salu explained to me, once again using the distinctive hand movement, downward-outward and inward-upward.
With Jocelyn's and Hessa's movements between Pororan and the Buka mainland, I am back at where I started: the constant coming and going that forms the background on which kin relations on Pororan appear. The background becomes foreground, and movements acquire form and significance, as people observing those movements ask those who come and go, or ask other observers, questions about their movements. They put their questions in those particular terms that will clarify the relation at issue, as well as the physical destination of the movement: Jocelyn was setting off from Lulutsi to "return" to Nova, her mother's place. She never "returned" to Lulutsi -to say so would have made no sense. (One of the small children at the hamlet once made this mistake and was ridiculed for days.)

MOVING AND CAUSING TO MOVE: MATRILINEAL KINSHIP AT SEA
But who or what was causing these movements, and what was keeping them on track? Audrey Richards' (1950) classic statement on matrilineal kinship among the Central Bantu resonates with the Pororans' statements about the matter. Depending on residence arrangements, matrilineal kinship, according to Richards, could be characterized by a "constant pullfather-pull-mother's-brother" (Richards 1950:208).
Pororan Islanders, too, say that a woman is pulled to his place by her husband, is made to settle down there, but is granted permission and even encouraged to return to her matrilineal group's place from time to time. She, and over time her children, are pulled there by the woman's brother, the children's mother's brother. He requests their return for mortuary and marriage rites, asks that they help with other ceremonial as well as everyday affairs, or simply wishes to see "my people". However, while Richards describes the relation between father and mother's brother in terms of a competition over a woman's "labour, her property, and her child-bearing powers, as well as rights over [his] children's work and their marriages" (ibid..), on Pororan, father and mother's brother described themselves as allies. In "pulling" a woman back and forth, they jointly brought about a movement that the woman performs, and in which her relation to each of them and their relation to each other becomes visible.
Of course, there are good and bad allies. Pororan brothers sometimes complained about their sister's husband, not because he took her away, but because he "held her tight" (holim em pas, Tok Pisin), that is, he wouldn't let her visit them (and they suspected he mistreated her and didn't want them to know). Pororan husbands, in turn, grumbled if their wife spent long periods away from their hamlet, on the grounds of visiting her matrilineal relatives (though they suspected that she might be "going around" elsewhere). Despite these complaints, the relations between brother and husband could hardly be described as a tug of war in the cases I know of. While Richard's description might evoke an image of a woman almost torn to pieces between her brother and her husband, an image closer to the islanders' complaints would be that of a movement, ideally smooth and regular, temporarily being halted or perhaps going off in random directions.
The Pororans emphasized that women's movements between their husband's and their brother's place never quite matched the ideal, even when both husband and brother behaved impeccably toward one another. First of all, women had a will of their own. They might "go missing on the way" of a 'return' to their brother, that is, change their mind and go somewhere other than planned, for instance, see friends along the way. This rendered the significance of their movement as a return to their mother all but invisible and "confused people", as Tsireh once complained about a niece who had been "all over the place" instead of "going straight" from her mainland relatives' place to her husband's. Moreover, people other than her husband and her brother might try to seize control of a woman's movement, jokingly or seriously. They may offtrack her, if only for a brief chat along the way, and sometimes, a new relation would emerge. For instance, marriages were often traced to encounters on the beach, where women see and like the fish that a man is bringing back from the sea. The islanders liked to dwell on the details of such encounters, moments of radical contingency whose relational outcomes hinged on the aptitude of those involved at turning the situation to their own ends. Central to the islanders' accounts of such encounters is the transformation of a visual impact (seeing someone, or seeing something one likes), first into desire (wanting it). Next, "turning the head" (tanim het, Tok Pisin) of the person visibly affects her movements, and if the alteration persists, her social relations (Schneider 2011).
Even the relations that the Pororans considered the closest and most enduring, relations between mothers and their children, were subject to change in this way. For instance, a woman I knew felt neglected as a child by her birth mother, who had many children to look after. At the age of four, she became attached to a paternal aunt, who was happy to take her in because she had no daughters of her own. The woman came to be known as the aunt's child. When the paternal aunt died, however, she made an effort to return to her birth mother. She made a point of visiting the birth mother daily and helping her around the house. Eventually, others on the island came to speak of her, not merely as her birth mother's child (as the woman hadn't been called for years), but as her 'good child'.
In 2005, this woman was about to become an "other mother" herself to her brother's son, three years old, whom her brother often left in her care when going to Buka Town for work, together with his wife. When the boy began to call his aunt 'mama', however, his paternal grandmother intervened. She was concerned that her daughter, who was pregnant, would be over-burdened after the birth of her baby if she had another child to look after. So one day, when the boy was clinging to his mama, his grandmother walked over, pulled him away and told him to 'go and look for your own mother'. The boy cried but obeyed. The next day, when his uncle went to Buka Town, he asked to be taken along to see his mother.
Although people could acquire other mothers besides their birth mothers by altering their habitual pattern of movements, the incorporation of women (not men) into another matrilineal group than that of her birth mother seemed to be rare. The Pororans argued that her relatives would not let women go lightly, because women 'continue the line'. 9) Women themselves, too, were stronger than men; they remembered where they came from, and they would always return. Some islanders used the term 'continue the line' in a manner that allows one to interpret it as 'continue the movement'. Others, however, interrupted themselves when discussing such matters with me and said that if I wanted to know more about this, I would have to ask their traditional leaders, or even better, their mainland relatives. 'The matrilineal groups do not stop here. They do not stop on the island', people said. The matrilineal groups, indeed, reached all across the Buka area, and all the way back into ancestral times and up the mountain of Punein, the place where all Buka people came from.

SHIFTING ATTENTION
The knowledge related to these geographically more far-flung and temporally deeper relations of matrilineal kinship was the privilege and burden of the traditional leaders (see also Sagir 2003;Sarei 1976). Male and female traditional leaders in Buka are formally introduced to their counterparts in other local branches of the matrilineal group, and are taken to the group's ceremonial houses and the places of its spirits on the mainland as teenagers. The traditional leaders currently active on Pororan had been taken to mainland relatives by their seniors, and they had learned on those visits how their groups were connected. Gradually, as their seniors grew old, the new generation grew into their role of engaging with their counterparts from other local branches of their matrilineal group, on behalf of their relatives on the island. This included, for instance, travelling to the mainland with baskets of fish in exchange for baskets of starch food for a feast at a later date, and attending mainland ceremonies. In 2004-05, it also included formal meetings for the purpose of re-establishing connections disrupted by the Bougainville Crisis and refreshing the leaders' memory of how they were related to particular local groups and people.
The terms in which kinship, and particularly being of 'one mother' were accounted for on such occasions differed fundamentally from those ordinarily used on the island. At two such meetings I joined, the traditional leaders present first listed place names along the group's migration route. They then recited the names of female ancestresses of each local group, up to the name of the ancestress who was the first of their group to settle in this location. Finally, they clarified the genealogical relation of this woman to the first ancestresses of other local groups. Bill Sagir (2003) has argued that genealogical proximity and priority of settlement as fundamental principles of reckoning relationships, on Buka as elsewhere in the Austronesian-speaking world. Sagir's argument seems fitting with respect to those occasions, 10) though not with respect to everyday accounts of kinship on Pororan. Ordinary Pororan Islanders listening in on the meetings of their traditional leaders found the careful discussions of the order of names in genealogies boring and the mere listings of place names to account for ancestral migration lacking in detail. How could they "be clear about this clan of ours", they asked, if only the names of places along their migration route were provided?
Traditional leaders can answer those questions, and can indeed switch back and forth between two modes of accounting for matrilineal kinship. Once, at an informal get-together of his (island) relatives after the visit of a mainland leader of his group, I had the privilege of listening to Tsireh responding to his relatives' questions. To the bare list of place names provided during the formal meeting with mainland leaders, he added details about ancestral movements. Talking in the first person singular, he explained which ancestor had come to Pororan, with whom (the ancestor of an allied group), that they had come on their own, found the place empty, and pulled other people there after they had settled. He merged his story of his ancestor's settlement of the island with stories he liked to tell of his own return visits to places along the ancestor's migration route on the mainland. Those stories were already familiar and perfectly comprehensible to his relatives. Upon being reminded of them, his relatives began adding to Tsireh's explanations of his ancestor's movements by extrapolating back from his own. While some remained unconvinced, other relatives affirmed that, although they didn't fully understand the mainland visitor's story, it might, indeed, be 'true'.
Rather than recount here how exactly Tsireh moulded the mainland counterpart's story into a shape that his saltwater relatives could appreciate, 11) I merely note his achievement as a source of inspiration for this paper. In the following section, I re-visit ethnographic accounts of kinship from landed settings in Melanesia (outside Buka) from a saltwater perspective. I highlight interests in movements that are conveyed in the original accounts, but that remain submerged under a focus on substance. The aim is to show that a richer understanding of kinship on land, too, is possible if we complement the common focus on substance with attention to movements learned from saltwater people.

MOVEMENTS AND KINSHIP ON LAND
It is worth returning to the original ethnographic problem that the focus on substance was meant to address: among PNG highlanders who affirmed a dogma of (patrilineal) descent, co-resident nonagnates were nevertheless readily incorporated into groups. While Langness (1964) set up an (literal) equation between kinship and coresidence, Andrew Strathern wanted to take this equation metaphorically and suggested to examine the understandings that underpin this metaphorical usage. The mediator between kinship and coresidence, he argued, is substance, specifically, food grown on clan land.
A saltwater-inspired re-reading of this original suggestion could take off at different levels of abstraction. At a fairly high level, one may ask for clarification regarding the direction of the metaphorical mediation performed by substance, and the mechanisms that determine this direction. 12) In A. Strathern's "static" notion of metaphor, substance seems to form the intersection of descent and coresidence: because both involve transfers of substance between people (but see Bamford 2004), people can become kin by coresidence. But how exactly does this mediation work, and why is descent a dogma and coresidence a practice, and not vice versa? What "pulls" the transformation that way? Roy Wagner (1974) has suggested that descent is an aspect of an ideological commitment to the existence of stable, clearly bounded groups shared by anthropologists and colonial officers, but not by PNG highlanders. If so, then the analysis of kinship as substance needs an ethnographically more appropriate centre of gravity. What might it be?
Adopting the Pororans' focus on movements, one cannot fail to notice the overwhelming evidence in the literature that substance does not "stop" at the boundaries of clan land. Pigs, fed with tubers grown on clan land, are passed on in ceremonial exchange (e.g. Feil 1984;A. Strathern 1971). Women move "between" clans (see esp. M. Strathern 1972), and valuables move against them (e.g. Wagner 1967). In these writings, the common question of how to "cut the network" (M. Strathern 1996) is subsumed under different theoretical and ethnographic agendas. By taking a saltwater perspective and asking explicitly about movements, this common concern can be foregrounded again.
A good starting point is one of the texts that A. Strathern used in support of his original argument about substance as a symbolic mediator between descent and locality. This is Salisbury's (1965) description of Siane rituals, in which the concern with stopping substance is readily apparent. Siane rituals are an important occasion for the transmission of clan substance, which includes spirit as well as corporeal substance. Salisbury concentrates on initiation rituals. In male initiation, boys are shown and hold the clan's flutes, and they take up the spirit of their clan in the form of flying-fox meat and pork. They also expel maternal substance from their bodies by letting blood. Salisbury highlights the contrast to the initiation rituals carried out for girls when they first menstruate and "give birth" to the clan spirit. They are given female substance (the meet of rat and opossum) to eat afterwards. When they leave the clan and move to their husband's land in marriage, they take female substance along, but the clan spirit stays behind. A. Strathern concludes: "Locality and descent are […] exactly fused" (A. Strathern 1973:31).
A saltwater-inspired reading would focus on spatial elements of Salisbury's description that do not make it into Strathern's analysis of substance as a symbolic mediator. The blood-letting during male initiation, as Salisbury recounts, takes place on the boundary of the clan land. The initiates are sent there on the pretext that they must rescue a war party of their clan that has been attacked by enemies. They come across the "injured" men at the boundary of the clan land. The injured men raise and bleed the boys' noses with reeds. The boys then return, decorated as warriors. Female initiates, by contrast, are confined on clan land while they are menstruating, and emerge, not from the boundary of the clan land but from the inside, from behind a screen inside their house. They travel to and beyond the boundary of the clan land later, at marriage, and are lost to the clan. This contrast between the physical movements of men and women coincides with a parallel movement of male and female substance. In the case of both male and female initiation (and marriage), female substance travels to and beyond the boundary of the clan land and disappears. Male substance, by contrast, is retained and embellished: the boys return from the clan boundary, where they have shed female substance, as warriors; the (clan) spirit "child" of the woman at menstruation is born in confinement in her house and remains there. In the case of the marriage prestation for the "oldest sister" of a generation, this retention of male substance is further highlighted by the gesture of giving pork (clan spirit) to the bride to take along to her husband's place, but then withdrawing it and giving it to a younger girl (who will stay).
While A. Strathern, focusing on substance, notes the coincidence of locality and descent, it seems equally possible to say that the Siane clan is constituted not by substance as such, but by movements: male substance (in the bodies of men and in meat) moves outward but, unlike female substance, is held up at the boundaries of the clan land and returns. This becomes clearly apparent if we compare the movements of male and female substance at initiation. We may then begin exploring what the movements of people and their gendered substances create. They create men (in male initiation), women (in female initiation), and drawing on James Leach's work on the Rai Coast, one may suspect that they also create places, by rendering salient the boundaries of clan land. I am not arguing against a focus on substance, but for complementing it with attention to movements. A comparative analysis of kinship that would focus on the constitution of clans through gendered movements would bring back into discussions of kinship many of the insightful analyses that exist, but that have been turned to different analytic ends. This includes the growing literature on gender, a large number of writings on exchange, but also historical studies (see esp. Wiessner and Tumu 1998) and research on land and land tenure. One question that would be interesting to explore with regard to the latter, and that emerges immediately from the above suggestions, is whether attention to movement can help us understand better what makes clan land and substance in the highlands apparently bounded and cohesive, by contrast with the assemblage of "rubbish", as the islanders say endearingly, that is Pororan Island.

CONCLUSION
My main aim in this paper has been to contribute to the limited amount of published "saltwater ethnography", with a focus on kinship because kinship is a core component of a lot of ethnographic research, certainly in New Guinea. "Landed" accounts of kinship often privilege substance, as an ethnographic focus and for analytic purposes. In most cases, such as the Siane case discussed above, this substance derives from the land and is bound up with it. This obviously raises the question of how the specific situation of saltwater people, with limited access to land, would affect their ways of engaging in kin relations.
Second, and with reference to my ethnographic field in particular, matrilineal kinship was a crucial element Bougainvillean identity in 2004-05, and it was enshrined as such in the constitution. Saltwater people shared relations of matrilineal kinship with others in the region, but they also admitted that they did not always understand the explanations that their mainland kin gave of those relations. Showing why this is the case, and exploring the systematicand systematically different -"saltwater" ways of perceiving and engaging in kin relations may help appreciate diversity, even in an area as small and tightly integrated as Buka. Finally, I have tried to indicate the contribution that a saltwater-inspired rereading of accounts of kinship on land, and specifically in the PNG highlands where many authoritative analyses of kinship in Melanesia come from, might have. A fuller exploration of this potential -through a closer engagement with the Siane material or another single case, with reference to a broader range of ethnographic examples, or ideally both -will be left to another occasion.

ENDNOTES
1) Ethnographic sources on this area include Blackwood 1935, n.d.;Rimoldi and Rimoldi 1992;Sagir 2003;Sarei 1976. 2) See, among others, Dove and Togolo 1974;Filer 1990Filer , 1992Kenema 2010;Mamak and Bedford 1974. 3) For details on the Bougainville peace process, see e.g. Braithwaite et al. 2010;Regan 2010. 4) See Blackwood n.d.; Parkinson 1907 for records of this trade in Buka and northern Bougainville, respectively. 5) Carrier 1985 noted that one kind of exchanged had replaced the other in the exchanges of Ponam Islanders with people on Manus. I found no evidence of such a trend in Buka in 2004-05. 6) For an early use of the term in New Guinea, see Glasse 1969:31, cited in A. Strathern 1973:32. 7) Tok Pisin is the lingua franca used in Bougainville and most other parts of Papua New Guinea among people who speak different local languages. Since many marriages on Pororan are mixed marriages, many children grow up learning Tok Pisin before they learn Hapororan, which is a dialect of Petats, one of the three languages spoken in Buka (all Austronesian). 8) A third, the oldest, was staying with his mother's oldest brother in another province. 9) The only case in which women changed her matrilineal group that I know of was in times of ancestral warfare. Captive girls were sometimes incorporated into the matrilineal group of their captor. They would later be married to a man of the same group, so as to "hold them tight". This relation is called "basket". In the case I know of, the woman's descendants were members of their new group on Pororan, but knew (although they did not visit) their ancestress' place and group of origin. 10) Sagir was specifically interested in rank. However, it seems reasonable to adopt his argument here. In accounts of migration and settlement, it is always both rank and traditional kinship that is negotiated, though the emphasis differs. The context in which Sagir worked, among a group of ambitious traditional leaders on the Buka mainland, could account for them and him privileging rank. At the meetings that I witnessed, when relatives of the same matrilineal group met formally for the first time after the Bougainville Crisis, their main interest was in refreshing their memories of matrilineal kinship between them. 11) See Schneider 2012:121-26. 12) For a style of analysis that does this, see Wagner 1972Wagner , 1986; see also Weiner 1988.