SPACE AND PERSON IN THE TROBRIANDS THE SELF AS LIVING AND DEAD1

The Trobriand language contains two spatial markers, o and wa, which designate types of space, space of the living and space of the dead. All geographic space in the Trobriand world is assigned into one or the other of these two categories. This essay delves into the hows and whys of this and illustrates why such a seemingly simple topic as space has proven to be so ethnographically and linguistically challenging

along this line at our peril too. Again, had I not happened to have been pushed by Kaduwagans, I would never have found this spatial usage. But, then, again, I was not looking for novel spatial categories. Now I know better.

The Beginning
When I arrived at Kaduwaga Village in the Trobriand Islands, I found myself having trouble using two words. I had learned from the dictionary provided by the Catholic mission that wa stood for 'in' and o for 'out. ' My problem was that sometimes the terms seem to switch places arbitrarily. Ana wabwala, means 'I go in (to the) house,' and Ana ovanu means 'I go out (to the) village.' But Ana ositor means 'I go out [?] (to the) store' when I go inside a store.
When I step onto a path I go wakeda, 'in (to the) path.' But, as with the store, when I enter a cleared garden plot, I go obaleku 'out [?] (to the) cleared garden plot.' Then it switches when I enter that same plot once it has been planted. I go wabagula 'in (to the) planted garden. ' I go okwadewa 'out (to the) beach.' I also go obwarita out (to the) ocean.' But then I still go obwarita 'out [?] (to the) ocean' when I enter the water for a swim in it. Still, if I get into a boat, I go wawaga 'in (to the) boat. ' To get totally confusing, missionaries have it that God is up there oraikaiwa 'out overhead,' but various Kaduwagans corrected me that this is "mistaken missionary talk;" it should be corrected to warakaiwa 'in overhead. ' My problem was serious for two reasons. First, although Kaduwagans were generally tolerant of my many speech errors, my misuse of these two spatial terms bothered them, such that they kept trying to correct me. Second, I misused them a great deal because they are ubiquitous in the language. Most spatial locations are labeled by them. As people kept correcting me, I just could not remember which one to use when.
I tried consulting the experts I had on hand, Rev. S. B. Fellows and Bronislaw Malinowski, as well as words I had copied down from a typed dictionary loaned to me by the Catholic mission at Gusaweta. 2 Fellows was an early Trobriand missionary who contributed a grammar of Trobriand to the governmental patrol reports back in 1902. I had made a copy of this on my way to the islands. Consulting it produced the comment that, "The preposition O (at or in) is sometimes sounded as Wa, but there seems to be no rule followed in the matter" (1902:176). This did not solve my problem, although it did tell me that Fellows did not view the two terms as having contrastive meaning, as did the anonymous author of the dictionary from the Catholic mission. While that author reported the one term standing for 'out' and the other for 'in,' I found them to be mysteriously switching meanings. Instead, Fellows viewed o as standing simultaneously for 'in' and 'at,' and sometimes having its pronunciation switched to wa for unknown reasons. Fellows did not mention either term standing for 'out.' So, in this sense, the two sources provided different takes on these terms. However each author, in his own way encountered unaccountable variant term substitution.
Malinowski, the famous Trobriand ethnographer, did not comment on these terms per se. So I consulted interlinear translations of Trobriand texts, primarily in the second volume of Coral Gardens and Their Magic (Malinowski 1965(Malinowski [1935), to see if they could give me a clue. There I found much the same sort of switching around. In his case both terms were given all three meanings, 'in, ' 'out,' and 'at,' along with 'to' and 'from.'

What's Going On?
In my confusion, I listened carefully to sentence construction and asked, both myself and others, about what instances of o usage had in common that set them apart from instances of wa usage.
I gradually learned that o and wa, unlike the English prepositions 'in, ' 'out,' 'at,' 'to,' or 'from,' do not relate the sentence's subject to a geographic space while working through a verb, like 'He goes to the store.' Instead o and wa are used to identify types of geographic space.
Trobriand divides all geographic space into one of two types, wa space and o, and Trobriand grammar mandates that most Trobriand spatial nouns append the wa or o label to tell the listener which type of space the noun indicates.
Thus the sentence Ana wabwala, which literally says 'I go wa-house,' does not say, 'I go in the house.' It says, 'I go wa-type-space-house.' Likewise, Ana ovanu, does not say, 'I go out (to) the village.' It says, 'I go o-type-space-village. ' Let us take this a bit further. A Trobriand speaker can put a pronoun into a geographical sentence in a manner that approximates English, e.g., Ama matoya Tauwema 'I come from Tauwema.' Matoya translates quite nicely as 'from,' so now the generated sentence structurally approximates an English sentence. However, in most speech contexts the Trobriand speaker would not put that 'from' into the sentence. For instance, the question, "Where do you come from?" is routinely Ambeya kuma?, literally, 'Where you come?,' not Ambeya metoya kuma? 'Where from you come?.' The standard answer is simply, Ama Tauwema 'I come Tauwema,' not Ama metoya Tauwema 'I come from Tauwema,' even though it is possible to construct the latter sentence in Trobriand.
Note also that there is no spatial locater, either wa or o, in the response Ama Tauwena. Nor would there be one in a comparable sentence like Ana Tauwema 'I go Tauwema,' even though, as Tauwema is the name of a local village, the comparable 'I go (to a) village,' would be Ana ovanu. Why not? As one of my informants put it, "Kaduwaga is the name of our village, but it also is the name you gave your cat back in 1971. A name can be attached to anything." This may be amateur linguistics, but the point is valid. Wa and o spatial-type indicators are prefixed onto geographic space nouns to tell the listener what type of space they denote. But formal names which those nouns might carry, like this village is called Kaduwaga, or this boat is called King of the Waves, or this establishment is called the Comfort Inn, are not themselves geographic space nouns. Therefore spatial-type indicators are dropped.
To summarize, the store interior is o space. The path is wa space. The cleared garden is o space. But that changes when it has been planted. Then it becomes wa space, like the house interior and the path.
The beach and the ocean are both o space, and the ocean remains o as I am swimming in it. It is not just o space as I am looking at or across it from shore. The boat interior joins the house, path, and planted garden as wa space. So does whatever space God is occupying as He is up there overhead. What, then, is the difference between o and wa?
I started asking, and my interlocutors' response was that o space is space that offers simatala whereas wa space is space is space that offers nigwanigwa. I already knew that the word nigwanigwa means difficulty, complexity, and/or confusion. I did not know the meaning of the word simatala.
I asked a couple of people to define the word, but their explanations did not seem helpful, so I started listening for speech uses. As I did that, my first two captures involved my hostess contrasting it with dudubila [dark]. The first time my hostess said, "Your room is dudubila. I'll get the lamp and bogwa bisimatala." The next time, she was on the house porch and called in to me, "Could you hand the battery [battery-run lamp]. It is dark here. With the battery bogwa bisimatala." In both instances I translated the bogwa bisimatala as 'then (bogwa) it will be light (bisimatala).' But as I tried to use simatala as 'light,' I found that I could not use it in sentences where we would often use the English word 'light' e.g., 'Hand me the light,' or 'The house has a lot of light this morning.' In the case of 'Hand me the light,' as I pointed at some light, like a flashlight, the person I was asking responded by asking me what I wanted. As I pointed again, my interlocutor would look around and name: 'the fire,' 'the lantern,' 'the torch,' 'the battery,' 'the benzene [cigarette lighter].' In the case of the house having a lot of light, my friends were at a loss, seeming not to understand what I was trying to say.
I listened further. My next clue came early one morning as I stood looking out to sea towards Tuma Island, which was hazed in. My hostess said, "This morning we can't see Tuma very well because it is very hazy. Later the haze will clear and bogwa bisimatala." Now simatala started to unfold. The word does not mean 'light' but something more like 'unobstructed or clear visibility.' Put a light into a dark area, removing the obstruction of darkness, and the space has simatala. Lift the haze between the would-be viewer and Tuma Island so that Tuma becomes visible, and Tuma has simatala.
From an etymological perspective this meaning makes perfect sense. The word consists of si-'existing,' mata 'eye,' and -la 'its (vision),' i.e., 'the eye's existing ability to see.' Thus o geographic space is space which offers unobstructed/clear vision, while wa space is space that does not.
In many instances, this interpretation seemed to make sense. Take something like a beach, which is an open visual area versus a grown garden, full of tall plants, which definitely is not. Or take a village center, a large clearing versus a narrow winding path also edged with tall garden or bush growth. The former is an open visual area and the latter is not. However, there are places where this analysis breaks down. A Trobriand house interior is not particularly visually cluttered space. Nor is a boat. How do they contrast with a store interior? It seemed to me that you can see across all of these equally well. So I asked some more.
The answer I received was that a boat or house interior is no more cluttered than the inside of a store. However a house or a boat interior is a much smaller space so that only a few people can gather together to see anything inside it. By contrast a sizeable number of people can gather together to see something at once inside a store. O space is space where numerous people can come together to see the same thing at the same time. Wa space is space where that cannot happen. Quite possibly more than two people can come and look about together, but not very many more.
As my friend, Kamsieboda said, "If you are in o space, many people could be with you and see the same thing you do. If you are in wa space, this cannot happen. The house is too small. The boat is too small. The path is too narrow. The garden plants are too many and too tall. You understand how it's not simatala." All right, I've got it. O space is space where numerous people can come together to see thing(s) simultaneously and wa space is space where numerous people cannot do that. But note that as Kamsieboda closes his statement, he expands the definition of the word simatala to incorporate the idea of simultaneous vision by numerous people. After explaining how wa space is space where numerous people cannot come together to see the same things simultaneously, he finishes his statement saying, "You understand how it's not simatala." Up to that point I understood simatala to denote unobstructed visibility. I did not realize that the term also requires that simultaneous unobstructed visibility exists for numerous people. Nor in fact does simatala require this. My hostess' statement that my lantern created simatala in my small wa space room was perfectly proper. The word can be applied in any instance of speaking of visual clarity versus lack thereof. Still, it also carries the idea of clear or unobstructed vision in situations where numerous people can see the same thing(s) simultaneously. As I followed up with others on Kamsieboda's statement, I discovered that people uniformly agreed with him and thought that when they told me that o space is space of simatala they were telling me that is was not only space of clear vision but space where numerous people could simultaneously see together.
The challenge clearly went beyond ordinary language learning. The critical term, simatala, turned out to be one I could get at with a bit of work, but it also turned out to have a secondary meaning that I would not easily imagine. In fact, many Trobriand words have meanings that are unfamiliar to Western speakers of the language, meanings that are unexpected and, therefore, difficult for those speakers to generate despite the endless patience of their local teachers. Meanwhile, those teachers are hampered by the fact that those meanings are obvious to them. That is the challenge I faced in trying to decipher the significance of o and wa, a quest that forced me to come to terms with the construction of the Trobriand person and magic.

So What?
Once I figured out the difference between o and wa space, I was faced with the question of why this particular distinction is important in the Trobriand world. Why does the language so insistently mark space, constantly reminding both speaker and listener that it is either space where numerous people can come together to see the same thing or space where they cannot? Why did it bother Kaduwagans when I got it wrong, such that they kept trying to correct me? I had no idea.
I started asking more about space beginning at home, where a handful of people would come by each evening for tea and cigarettes. Thus we had a handy spatial venue and time to chat.
As we chatted, I discovered that a house interior, which is wa space, is space of the dead. Initially, then, I imagined that a house interior is somehow dedicated to ancestral family members. I had read about Asian societies in which people have shrines to the family dead inside their houses, or the Tikopian practice of burying the dead under the floors of their houses (Firth 1936). But what about other wa space? Is it space of the dead too? I asked, and sure enough it is. All wa space is not only visually non-aggregative space, it is also space of the dead. This does not mean that Trobrianders also dedicate paths and planted garden areas to their ancestors. Nor do they dedicate any part of their house interiors to their ancestors. So in what way is wa space the space of the dead? The answer requires a look at Trobriand cosmology, starting with their traditional view of the earth itself, and then of the nature of the person.
The traditional view of the earth holds that it is something like thick disc or a layer cake (see Figure 1). Its upper surface is solid ground. This is where (embodied) living people, tomota, reside. This solid surface gradually thins down to a watery middle, which in turn thins down to an amorphous/gaseous underside.
[Disembodied] dead people, baloma, reside on the underside. The solid upper side, our side is o space, and the amorphous lower side is wa space.
So there is a disk with a solid upper side, a watery middle and gaseous lower side; the top side is the land of the living and o space, while the lower side is the land of the dead and wa space. The middle-indeed everything below the upper surface-goes along with the dead as also wa space. Commonly, however, informants simply draw the contrast between the o space of the top and the wa space of the bottom. That is to say, as Kaduwagans describe the earth, they start out by describing it as a two-sided disk, solid on top, Now let us look at the solid upper side, the living people side (see Figure 2). It is o space, in contrast to the lower side and interior. But it is not just o space. Rather, it is o space punctuated with areas of wa space, such as house interiors, boat interiors, paths, and planted gardens. And although the earth's upper surface, basically o space, is where the living reside, the wa space that punctuates it is said to be space of the dead, just like the wa space of the earth's underside where the dead literally reside.
To understand why this is so, we have to move to the cosmological construction of the person. Here we come down to a fundamental difference between Trobrianders and Europeans. The reason wa space exists on the top side of the earth is that, unlike living Westerners, living Trobrianders are simultaneously two things. They are at once the living and the dead.

The Living/Dead
To summarize thus far: embodied living people, tomota, reside on the earth's solid upper side, and disembodied dead people, baloma, reside on the gaseous underside. Also, paradoxically, while the dead are not living, the living are both living and dead.
According to Trobriand cosmology, a dead person is a baloma. The easiest way for a Westerner to imagine what a baloma is like is to think of the Christian God because Kaduwagans insist that the Christian God is a baloma, just as, when dead, each of them is a baloma. An embodied living person, tomota, is a combination of a wowo 'body' and a waiwaia. A waiwaia is a combination of a nano 'cognitive mind' and a baloma. So while dead disembodied people are not alive, being only baloma, the living are both alive and dead, being wowo, nano, and baloma.  Malinowski (1929:170-179) observed that living Trobrianders start out as baloma 'spirits of the dead,' perhaps better put as 'disembodied minds,' who regress themselves into waiwaia 'infant spirits,' perhaps better put as 'infant disembodied minds.' Waiwaia come up from earth's underside to the earth's surface through various holes in the ground, enter into women's wombs, acquire surrounding bodies, and enter into life. For a limited time they then form the core of the living person. When their surrounding bodies die, they return down a hole in the ground to the earth's underside and resume existence as baloma.
Kaduwagans do not agree with Malinowski that when a baloma moves to enter into life, it wipes out all of its knowledge and regresses into a know-nothing "infant spirit," i.e., a waiwaia. Instead, they, along with other Trobrianders I have asked from other villages both on Kaileuna and Kiriwina Islands, say that each baloma which desires to enter life creates and attaches to itself a new, temporary form of cognition, nano, which is the form it uses to take cognitive precedence so long as it resides inside a body. This cognitive entity starts out life entirely empty of knowledge. It is also limited to picking up knowledge through the body's sensory inputs. The baloma which has generated the nano and to which the nano remains attached functions essentially like the Western subconscious or unconscious mind. For instance, the baloma is responsible for making one's body move when one wants it to, for making one fall asleep and wake up, for making one feel hungry, for making words come out of one's mouth: in other words, for things the body does that one does not consciously control. And it is this combination unit, the baloma plus its nano that constitutes a waiwaia.
The Trobriand term nano has been routinely translated into English as 'mind' (cf. Senft 1986). This translation is consistent with such constructions as Nanogu ateya 'I change my mind,' and Nanogu itaboda 'My mind is blocked/shut.' But it became confusing to Trobriand speakers when I tried to use it to speak, for example, of things like nanona Yaubada 'the Mind of God' or nano isibogwa taga gala tagisi 'mind that's there but we can't see/perceive' i.e., 'unconscious mind.' Those things are not covered by nano, so people had to struggle to figure out what I was trying to get at. I was applying nano as if it carried meanings that are carried by the English word "mind," but nano turns out not to carry such significations. As with simatala, it took time for me to understand what nano meant.
I noticed that perhaps Annette Weiner encountered some similar problems. Such a conclusion is suggested by the fact that she neither supported Malinowski's regression picture on waiwaia, nor did she create any sort of specific alternative. She simply left the matter at defining a waiwaia as "The spirit child that is thought to come from a baloma and is thought to cause conception" (Weiner 1988:171). Nancy Munn, in her ethnography, The Fame of Gawa, provided some help. She writes: Sleep, fainting [-kaburamata], and death [-mata] have certain parallels. In the latter two states, the balouma spirit or life essence that is inside [wa-nuwa-ra] the body becomes separated from it and may move [-rarora] in the external world unseen by others, whereas the body itself is motionless or lifeless. Although the separation of balouma and body is not a necessary concomitant of sleep, it may also occur during sleep. Thus the balouma of a witch may be moving around in the bush when the witch's visible body remains asleep. (1986:75) Kaduwagans said the same thing. During sleep a person's baloma can leave the body and venture out, having experiences and adventures as it wanders. Sometimes these are remembered by the conscious part of the mind, like dreams, and sometimes what are remembered are actually dreams that the conscious part of the mind has produced. Upon reading Munn's passage I remembered seeing a woman go into trance. I saw her lie down on the floor, and her baloma left her body and went off to the world of the dead for a visit. Other women stood guard over her body, inert but alive during the temporary absence, so that it would not be disturbed while her baloma was gone, as any shock might render her unable to cope with an immediate bodily crisis. I also remembered that one should never shake a sleeping Trobriander to awaken him or her. The person's baloma might be off wandering and not able to get back quickly enough, even though it only takes a split second, to defend the body from harm.
Armed with these pieces of information I was able to affirm that nano is the added-on mental category that creates the waiwaia composite. More recently, Mark Mosko has indicated that, when speaking of disembodied baloma as mind and thought, presumably conscious, his informants use the terms kaikwabu and peula. He does say in one line that, like the living they have nano and nanamsa, but stresses the other usage, indicating that, ultimately, what gives all the bilu baloma 'baloma forms,'3 their commonality, is their possession of kaikwabu and peula (Mosko 2014:23). My informants said the same thing, the only exception being that they would quibble with this one line where he says that disembodied baloma have nano and nanamsa. That, no, that is an inappropriate language usage. That it is more appropriate to say that all instantiations of baloma are kaikwabu (conscious) 'mind' with peula (conscious) 'thought' although, when they have waiwai-ed and entered a human body, they fluxuate in this, reverting to the kwaikwabu/peula status whenever they leave it and returning to the nano/nanamsa status whenever they re-enter it.
Part of the problem here, if it is at all correct to say that there is a problem, is that anthropologists have tended to view the baloma within each person as the person's spirit or soul, in a somewhat Christian-like manner, rather than viewing the person as inherently being a baloma. This despite the fact that, since Malinowski's initial reports, it has been known that the living person begins as the baloma transformed into waiwaia. But for us, the soul is the soul, and the mind is the mind. Therefore, it has been difficult for anthropologists, myself included, to articulate Trobriand concepts with exactitude.

Again So What?
In a way, the Trobriander is much like the Western person: someone with a living body, a conscious cognitive mind, and a subconscious-unconscious mind which otherwise runs the body. Yet, in another way, the Trobriander is quite different from the Western person, because the subconscious-unconscious mind is effectively the Mind of God.
As Trobrianders render Yaubada as a baloma, they are fully aware of how Christian missionaries characterize God, as all-perceiving, all-knowing, and all-powerful (i.e., capable of manipulating any and all force in the entire universe at will). Living Trobrianders think and perceive, like Europeans, through their nano, their conscious cognitive minds. Nonetheless, those minds are backed by much more powerful subconscious-unconscious minds.
One consequence is that the nano can engage in two alternative modes of agency. It can take action either by calling on the body to perform physical labor to meet its needs and desires or by calling on its partnering baloma to expend directive energy/force to alter things in the external environment. Insofar as the Trobriand person does the former, s/ he is acting like a Westerner. In Trobriand terms, s/he is engaging in paisewa 'work,' the mode of production that is appropriate to life. Insofar as the Trobriand person does the latter, the Trobriand person is engaging in what Westerners are apt to dismiss as a fanciful enterprise, beyond the capabilities of a living person. In Trobriand terms, s/he is engaging in meguva, [magic], the mode of production appropriate to death. Meguva is inappropriate to life because its practice confers selective coercive abilities on certain individuals, thereby robbing others of the ability to live fully. A standard disembodied baloma's formula for action can be described as migi 'face,' magi 'desire,' and migai 'transforming action.' 'Face' here is short-hand for perception. The baloma turns its "face" around to perceive what exists. Then it formulates a desire. Then it issues a command which simultaneously releases power to effect that command to alter the surrounding environment. Think of the Biblical assertion of God perceiving darkness, deciding that He wanted light, commanding, "Let there be light, and instantly there was light. In a monotheistic system, such an all-powerful God is logically straightforward. But what if there are many gods, and they want different things? Do they tread on one another's toes? From the Kaduwagan perspective it does not matter because, as each of them is all-powerful, each can each take care of itself. Each can stand up to the others and defend its interests. As Ebouna put it, "If you are all one another's knowledge and power equals, no one can push on the other." However the situation is different among the living. A person who wants to tap into his/her baloma to get it to migai, i.e., to produce a transforming action, has to know how to do this. One cannot either consciously think or physically voice something like, "Hey baloma, tell the wind to change direction!" and have anything happen. Instead one has to know a special formula and utter it in a special way. That can bridge the barrier between the nano and the baloma and spur the baloma into action.
The person who would have the baloma take action must know the special formula for each endeavor. These formulae are esoteric and valuable, passed along family lines as well as bought and sold in the village marketplace. They are not all equally available to everyone, and the uneven possession of magic confers greater personal power on some people than on others. For example, suppose that Sam has meguva that will get his baloma to attract sea bass and Joe has meguva that will get his baloma to attract local cropensuring rain. Who is the more powerful person? Obviously, Joe. Furthermore, Joe's magic is not only beneficial to him; it is socially beneficial in that it can ensure adequate water for other people's crops. But it also can give Joe coercive power over local people because he can threaten to use it inappropriately to flood those crops and damage or destroy them if their owners act in a way that displeases him.
To put it succinctly meguva is the language of coercion. My host, 'father,' and friend Paramount Chief Katubai commented, "It would be nice to be able to be a boss like the Western bosses and have people work at my direction for money instead of because they are afraid of me." I said, "Why are they afraid of you?" He replied, "Armed with my magic, I'll kill them if they disobey me." Of course, paradoxically, Katubai's magic is precisely why he is paramount chief. Indeed it is why the office of paramount chief exists. 4 When Kaduwagans speak about the opposite mode of production, the mode wherein nano remains true to itself and works with its surrounding body to pursue its goals by engaging in physical labor to produce certain items and obtain the rest through exchange, they stress voluntarism. I have been told countless times that no one is ever forced into any exchange. One is always free to decline requests without fear of negative repercussions. Likewise one is always free to decline offers from others. Additionally, one should always be grateful when the person one approaches agrees to an exchange, even going so far as to express this in funerals, as illustrated in this conversation with Katubai: Montague: "If a man were to die from whom I had purchased a coconut tree, would I give something at his funeral?" Katubai: "Yes." Montague: "But I paid him for the tree, and he is not one of my relatives." Katubai: "But you wanted the tree, and he did not have to agree to sell it to you. He could have kept it or given or sold it to someone else. But when you wanted something from him and approached him, that man held out his hand to you instead of refusing you. So you should stand up and pay something into his funeral. Not a lot, but something, to show your gratitude."

The Trobriand Person and O and Wa Space
Trobrianders draw a sharp distinction between the "social" person, i.e., the person who functions on the basis of nano 'conscious mind' as s/he interacts with consociates and is concerned with voluntarism, and what I will call the "anti-social" person, the person who functions on the basis of baloma 'unconscious mind' and is, thereby, willing to countenance coercion. As Weiner (1988:67-71) points out, at adolescence both boys and girl foray into the world of magic as they move into the realm of sexuality and sexual attraction. Moreover, she observes, "Attracting lovers is not a frivolous, adolescent pastime. It is the first step toward entering the adult world of strategies, where the line between influencing others while not allowing others to gain control of oneself must be carefully learned" (Weiner 1988:71). And again, "To influence or even try to control another person is difficult, yet such efforts are a major preoccupation throughout each Trobriander's life " (1988:66). The point is to be able to coerce without being vulnerable to coercion oneself.
In short, there is no Trobriand Islander who does not practice magic. Whether it is love magic, beauty magic, weather magic, the practice of magic is a routine part of Trobriand life. Indeed, the distinctive feature of a paramount chief is precisely his magic. But since the practice of magic is anti-social, no Trobriand Islander can afford to be seen to practice magic. And this is where the distinction between o and wa space is so crucial.

O and Wa Space
As my informants indicated, o space is space of simatala, space where numerous people can come together to see the same things simultaneously. These congregated people constitute a social body. What they simultaneously see whenever they gather are individual congregants who present themselves, both visually and verbally, as people who function according to paisewa, physically working to make things and engaging with others according to principles of voluntary exchange. When under public scrutiny, everyone typically denies possessing meguva, which would be perceived as coercive and, thereby, antisocial. To proclaim control of magic in a public setting is to invite a potentially lethal attack on oneself. After all, even a deadly attack on someone who acknowledges highly anti-social tendencies can be understood as a public service.
However, this scenario shifts once a person moves into wa space. There the person can, and does, reveal the "dead" part of the self, the part that has and uses meguva and which is given over to coercion. The audience in wa space never is sufficiently large that any of its members can retaliate without being accused of unjustifiable violence. Even if the would-be attacker tells others what s/he saw, for them it is hearsay. Even if the attacker is backed by a few others who also were present, it is hearsay to those who were not. To use lethal force under such conditions is murder, not socially sanctioned execution, and the attacker risks retaliation.
This holds even if many others see the same thing as the would-be killer but they are strung out over time in wa space. By revealing one's control of magic to a small number of people on one occasion, to another small set of people on another occasion, and to a further small set of people on yet another, it is possible to gradually reveal the self as a person of meguva to virtually everyone around and remain safe because the revelation stays anchored in hearsay. It is only more than that within each small set of people who were simultaneous witnesses.
To return to the opening question, what makes wa space the space of the dead is that it is space where the living can, with impunity, present their "dead" selves. While the "dead" self cannot be revealed in o space, there is no corresponding ban on revelation of the "living" self in wa space, so the more complete person routinely emerges there on self-presentation.
Weiner appears to cite a couple of instances where people do reveal themselves as magicians in o space. These instances are worth examining.
In the first case (Weiner 1988:40), paramount chief Vanoi was in a trade store on Kiriwina Island and declared that he controlled powerful magic. As has been discussed herein, a store interior is o space. Vanoi could do this because there were not enough people in the store to mount any physical threat to him. It is not literally the type of space that governs or constrains presentation of self. If only a few people are present in a given space at a given moment, a person may act as if s/he were in wa space. In short, the space may be such as to provide simatala, but if too few people are present to take advantage of it, the fact of being in o space makes no difference.
In the second instance (Weiner 1976:70--71) men stood up in the center of Kwaibwaga Village and held a magic-recitation contest. But, as Weiner indicated, this was on a very dark night. No one could see who was speaking; people could only hear the voices ringing out. Thus the rules of hearsay held.
This dark-night sort of ploy was also used by paramount chief Katubai whenever he found it suitable in his role as chief to issue any orders to Kaduwaga Village. After all, orders are coercive, deriving their force from magical prowess. Katubai rarely issued orders to Kaduwaga. But when he did, he stood out in the village center under cover of total darkness (never on a moonlit night). His voice rang out from there as he articulated his command (and corresponding threat to anyone who disobeyed).
I recall one night when he commanded everyone to turn out the next day to work on the fence surrounding the large field that was to be planted for the ensuing year. Work was lagging a bit, and he was concerned. The next morning I was walking by one of the neighboring houses, and the residents had packed up lunch and all and were heading out with their children. I stopped and said, "Oh, you are off to work on the fence." The husband said, "No. It's such a lovely day, and we haven't had any time off together in so long. We thought we'd take the kids and go to the next cove for a picnic on the beach." I was surprised and said, "But Katubai said that he would banish anyone from the village who didn't get out there and work on the fence today!" My neighbor laughed and said, "But maybe that wasn't Katubai. You couldn't see him, could you?"

Finishing the Terms
A newly planted garden, bagula, is o space, just as is baleku, the 'unplanted garden.' It only becomes wabagula when the plants get about knee high and it becomes unlikely that many people will congregate there.
Reboaga ("swamp") and oduna, ("bush") are both o, which may seem odd given that they both are full of tall thick growth. But Kaduwagans characterize these o usages as based on contiguity. A similar idea is expressed in the sentence, Ana obwala 'I go adjacent-space-house.' Adjacent-space is always o, whether it literally is o or wa space. In these instances it is literally wa space, because it is a path, and paths are wa space. The point, according to Kaduwagans, is that one only goes past or by, or even-in our English sense-through, swamp or bush while on a path. One does not really go in them. So if you are casually saying that you are "going swamp/bush," you use the adjacent-space o rather than their literal spatial type wa. You can use the wa if you mean to indicate that someone intends to go off a path into them. However, one does not hear that locution used much.
Nonetheless, oduna is unusual in that it routinely is said onaoduna with the na infixed. None of the linguists or ethnographers have any explanations for that, and it does not appear anywhere else. My interpretation has been that, as usually a na noun prefix is a third person possessive, as in his/her/its something, this is a statement that the bush belongs to some third person. If so, who would that third person be? My guess has been a tokwai, bush spirit, but this is not clear. So I asked several Kaduwagans. None of them had ever thought about this or even noticed the unusual na stuck in there between o and oduna. However, when I pointed it out, they all were quick to recognize it might be a possessive and noted that, if so, it probably is because the bush belongs to a tokwai. Here another old friend, Boyomu, gave a nicely typical closing statement: "It must be from way back when, so who can really tell, but that does at least make sense." O and warakaiwa come from the fact that God, Yaubada, is a baloma like all the others, and as such resides in dead space. He may be up there overhead, like the missionaries claim, rather than under the ground. However, He himself is invisible to the aggregated living and, therefore, so is his space. The missionary mistake, from the Kaduwagan perspective, lies in the fact that while people can all look up there and see the clouds and sun, moon and, stars, they cannot see God in His literal space any more than they can see other baloma in theirs.
Finally, although this article has focused on the meaning of the Trobriand terms o and wa, it really is only a start at that. It explains how they are used in the context of space that people can occupy. That is what all these spaces have in common, and it was this contextual usage that bothered my Trobriand interlocutors. Nonetheless, the Trobriand language does not limit the usage of either o or wa to this context, nor does it limit the terms to other contexts where they function as a substitutive pair. Neither does it limit them to other spatial contexts. They are mostly used in spatial contexts, but Trobriand also uses o temporally. The Trobriand linguist Gunther Senft (2006) has explored the o/ wa pair in some detail with respect to body parts versus space.

Conclusion
In closing, I return to the beginning. What started as my neighbors badgering me when I persistently attached the wrong spatial indicator to space-related nouns became a lengthy odyssey into language, space, and person. I am still unsure whether, through it, I learned more about the challenges of language acquisition, space, or people. What is clear is that people create spatial renderings that depend on their constructions of themselves as people and, as evidenced here, that they do that in ways which can escape years of excellent ethnographic attention. Our theories are inevitably limited by the data they purport to explain. Therefore, we must strive to search beyond our imaginations and report whenever we discover the previously unimagined. ____________________ considerable assistance and for helping me access the Depth files, as well as Richard Feinberg for his advice and editorial work on the chapter. 2 In his dictionary, recently available online, the Reverend Bernard Baldwin indicates that wa is a variant form of o, meaning 'in, ' 'on,' 'at,' 'by,' and 'with' (Bernard 2015). In his online dictionary, the Trobriand ethnographer Edwin Hutchins asserts that "the locative prepositions O and Wa are complete synonyms. Wa is used when the noun (or noun phrase) that follows has stress on the first syllable. O is used otherwise" (Hutchins 2015). However he also notes that there are exceptions, which only makes sense given that words like vanu and odila and reboag, which routinely take o, all have the stress on the first syllable.
Possibly Hutchins picked up this idea from the Trobriand linguist Ralph Lawton, who, in his Topics In The Description Of Kiriwina, says that "the preclitic wa is used where the following word bears primary stress on its initial syllable, the prefix o-is generally used elsewhere." He too adds, "However this needs further working.
Hutchins simply lists the two as locative prepositions with both meaning 'in,' 'on,' 'at,' and 'into' (Hutchins n.d., no pages). Lawton ranges further in his own dictionaries, listing wa as 'on, ' 'in,' 'to,' 'towards,' 'at,' 'near,' 'by,' 'with' and 'by means of' (Lawton 1993:332;and nd1: wwww1), and o as on, with, and at (Lawton nd2:OO PP RR 1). In the dictionary portion of his Kivila: The Language of the Trobriand Islands, Gunther Senft reports that both o and wa, which he renders as va, given Tauwema Village pronunciation, mean 'to' and 'into.' He adds that va is not used in connection with the names of villages or towns, but does not make any comment about this with regards to o (Senft 1986:342, 400).
In a later article, "Prolegomena to a Kilivila Gramar of Space," Senft (2006) asserts that if Kilivila speakers want to refer to a goal or location with a specific term, or they want to refer to a specified place at the destination of a motion event, they use the locative o. If they want to refer to the goal or location with its most general term, or if they want to refer to an unspecified place at the destination of a motion event, they use the locative va. He provides an example: Bala o buyagu, I-Fut-go Loc garden, 'I will go to the garden' (i.e., my personal, specific garden plot); Bala va bagula, I-Fut-go Dir garden, 'I will go to the garden' (general, unspecified expression for 'garden'). This example is a bit puzzling and possibly is a simple mistake. In his dictionary, Senft lists baleku, not bagula, as the term for a personal garden. All of the Trobrianders with whom I have spoken about the matter say that buyagu is the word for a village's gardens in general. Both baleku and bagula refer to individual garden plots, which in turn are possessed by individual people, baleku being unplanted and bagula being planted. In any case, he is empirically incorrect, as people routinely say Bala o buyagu and not Bala wa buyagu. This would be translated by English speakers as something like 'I'm heading out to the gardens,' a quite generic distance goal statement. 3 Being all powerful, baloma can take on any forms they want. While at any given time most baloma are thought to be simply hanging out in space, various of them are also thought to be instantiating themselves otherwise as things like various village spirits, tree spirits, warrior spirits, bird-sharks, and others. 4 Mosko (2014) argues that magical formulae call on disembodied baloma to perform. He does this despite noting that all earlier anthropologists have basically thought otherwise; that magical formulae at least mostly unleash powers from the magicians themselves, even if they did not know how this works. My own guess is that Mosko, a newer anthropologist on the scene, has been told the latter since he arrived after the Papua New Guinean government passed a national law outlawing magic and sorcery in the country and attaching heavy penalties to its practice. Trobrianders are very intelligent and sophisticated as well as having members in the national parliament, and came up with the idea that, as Yaubada, the Christian God, is a baloma and there is nothing illegal about saying any formal prayers to him to get him to do one's bidding, there cannot be anything illegal about saying any other formal prayers to any of the other dead baloma to get them to do one's bidding. So, recast this way, magic cannot be illegal.
Nonetheless, on my last fieldtrip, as I was about to walk close to a house outside of Losuia, I was warned back by a group of people because a man or men inside were in the process of performing powerful weather magic and the forces emaninating outwards from them might cut through my female pwasa 'soft-squishy' bodily substance and damage it. Stay back until they had finished. So I did.
My only quibble with Mosko is that he currently wants it to be that Trobriand magic calls on dead baloma to perform, when I and all other older anthropologists have been told otherwise. My own informants indicate that this is what Trobrianders are now telling outsiders because the Papua New Guinean government has outlawed magic and sorcery in the country, with hefty legal prison penalties. But as Yaubada (the Christian God) is simply one of the baloma and it is perfectly legal to make ritualized verbal prayers to him, it cannot be illegal to make ritualized verbal prayers to any of the others. So they are claiming that this is what magical chants are about. Even so, in 2014, I and other women were warned away from a house in Losuia where some men were performing powerful weather magic because the energy beams emanating from them could cut through and damage our soft-squishy (pwasa) bodies. If you are a smart woman you never get near any man or men who is doing this sort of stuff and tapping into his own baloma for this sort of energy release. It may get out through his hard body all right without harming him, but it is not safe for women.