Communicated into being: Systems theory and the shifting of ontological status

Dimensions of animation and personhood

Personification and animation are processes that address dimensions ranging from inanimate beings to animate ones, from non-communicative beings to full persons, from beings that are foreign and distant to members of a community. At the one extreme are inert objects without life or intentionality. I do not assume that this is their naturally given state and that only their animation is a cultural achievement. The conditions of object status need analysis like any other. At the other end of the range are animated beings, person-like communicators or full-fledged socialized persons who are able to live as members of human collectives, have kinship relations, permanent marriages, and so on. As with objects, definitions of personhood are specific to social and cultural context, but I assume that some notion of complexity, communicative abilities, social presence and responsiveness are always part of them. In between objects and persons are life-forces or potencies that are responsive to ritual manipulation, as found in many parts of Southeast Asia (e.g. Anderson, 1972; Fox, 1987; Kirsch, 1973; Tooker, 2012). Closer to full persons are persons whose integration into a specific human sociality is somewhat restricted, e.g. regarding accountability, marriage or language. Categories of such graded persons sometimes include strangers, children, severely ill persons, or slaves, and are continuous with some non-human persons, spirits or animals, as found among Native North Americans (Hallowell, 1960) or in Amazonia (Viveiros de Castro, 1998). These are often full persons in their own communities, to which living human beings have only limited access (Sprenger, 2014). This notion of personhood is inspired by the relational and dividual person, initially analyzed in India and Melanesia (e.g. De Coppet, 1995; Marriott, 1976; Mosko, 2010; Strathern, 1988; Wagner, 1991).

The problem I want to address specifically is one common to the study of Southeast Asia. In the cosmologies of this region, numerous non-human beings do not occupy stable positions in regard to animation or personhood. In the same cultural setting, but dependent on context, they variously appear as vaguely defined classes without discernible individuals, or as impersonal life-forces or potencies, while in others they emerge as communicative, well-defined, albeit graded persons. A well-known example is the notion sumangat among Malay (Endicott, 1970), which is variously personal soul, spirit guardian or malleable, impersonal potency. Lulik in Timor (Bovensiepen, 2014), ruwai among Chewong of the Malay Peninsula (Howell, 1984), or lennawa among Ifugao on the Philippines (Remme, 2016; see also Sprenger, 2016a) have comparable features. An example from Siberia is animal spirits that are sometimes concrete single persons and sometimes a generic class (Willerslev, 2004). Such shifts in personhood or animation have been mostly discussed in terms of accompanying shifts in modes of production or growing social distance between humans and non-humans, for example through the introduction of hierarchies (e.g. Descola, 2011: ch. 15; Hribal, 2007; Ingold, 2000: ch. 4; Knight, 2012; Naveh and Bird-David, 2014; Tapper, 1988). These studies assume that the ambivalence of beings is a function of historical transitions. While my discussion is relevant to these cases, I deliberately restrict it to the question of how the status of beings shifts within a certain sociality. Even in cases without relevant historical change, it appears that animation and personhood of non-humans are more or less elaborate according to the situation. This oscillation between emergence and retraction of personhood is the starting point for the following reflections. I am not going to evaluate the various theories of animation here (see e.g. Gayon, 2010) but focus on the potentials of a single approach, derived from the work of Niklas Luhmann.

The shifts of ontological status found in various ethnographic contexts unfold along different axes – between non-animation and animation, between potency and personhood, between graded person and full person. Each of these axes is defined in culturally specific terms and practices, ‘personhood’ being as specific as ‘objecthood’. However, in order to clarify the problem I want to address, I propose to abstract from these cases and tentatively address them by a single model, linking human persons, non-human persons, potencies, living and inanimate beings. This model I call the person-animate-inanimate axis, or PAI axis (see Figure 1). What is important about it is not so much the specific categories on the line, but the potential to move along it – not a typology of beings, but an analyses of processes. OPEN IN VIEWER

On this axis, personhood is the point of reference. Animation, inanimateness and personhood first of all describe different forms and degrees of impact on communication. Beings that are alive have a different presence in systems of communication than those that are not. Personhood, defined in terms of ability to communicate, accountability for doing so and the assumption of social roles, is the mode of full presence of beings – human and non-human – in such systems. Communicative systems assign to persons the potential to engage in communication in a durable, expectable manner (see below). In this respect, inert objects are socially dead, even when they are subject to communication. The PAI axis thus describes shifting distances from personhood. Each position on the axis is the result of recognition and communication with or about beings.

In order to elaborate this central idea, I will first employ concepts from Niklas Luhmann’s theory of autopoietic social systems, and, secondly, give an ethnographic example of the differentiation of spirits in an animist context. Finally, I will return to the question of animation.

A Luhmannian approach

The conditions of ontological shifts and the terms and typologies associated with the beings involved vary according to the setting. Thus, a general typology of non-humans must come as the result of extensive comparison in the manner of Descola (2011). But for analyzing personhood and animation as communicative processes, starting with a typology of societies, cultures, ontologies, or even contexts and beings would be too restrictive. Both human and non-human beings move between ontological states quite easily. Thus, even the identification of contexts should be considered as a result, not a condition, of the analysis. What we need first are flexible theoretical instruments that draw attention to the processes by which these movements come about. I suggest that Niklas Luhmann’s theory of social systems provides such devices.

First, two caveats. Bruno Latour is usually credited with the introduction of non-human beings as actants in current social theory. These actants are involved in networks in which their specific qualities emerge in interaction between humans and non-humans. They impact interaction as agents, even without intentionality. ² For Latour, the very existence, the physicality of actants, is sufficient for having an impact. Therefore, networks by definition encompass very heterogeneous beings (Latour, 2008).

Niklas Luhmann pursues a very different approach. For him, the unity of a system is based on the unity of its elements. Social systems consist of communications, and nothing else. Bodies and any physical qualities are in the environment of social systems. Thus, while Latour’s networks integrate all kinds of beings into a mutually constituted existence, Luhmann distinguishes systems quite rigidly along a modern-type ontology that corresponds with the differentiation of mind, body and society (see below). Social systems made up of communications, psychic systems made up of thoughts, and organisms are entirely separate (Luhmann, 1984: ch. 7, I; Luhmann, 2008). ³

This has led to the common assumption that Latour and Luhmann do not go together well, and at least Latour has been outspoken in his criticism of Luhmann (Latour, 2005: 270, 414, fn. 38 ). ⁴ Indeed, the principles on which these approaches are built seem entirely exclusive to each other. While Luhmann stresses differentiation, Latour, and even more so Ingold, aim at hybridity and merging. While Latour points at the empirical mixture of human and non-human agency, society and nature, Luhmann stresses that communication – and thus science, technology, etc. – is only possible through its self-referential closure (Reddig, 2006). However, there are a number of parallels between the two theories, which have been pointed out before, in particular regarding personhood (e.g. Belliger and Krieger, 2006: 32–7; Teubner, 2006). ⁵ Among them is the insight that sociality results from a continuous and highly contingent step-by-step process that ‘makes’ its actors (e.g. Latour, 2008: 156; Luhmann, 1984, ch. 1, III).

The second caveat concerns Luhmann’s relation to anthropology. While he is taking classical anthropological literature into account, his theory is self-consciously part of and about modernity. The few anthropologists who have used his concepts have predominantly applied them accordingly, to modern-type institutions and organizations (Gershon, 2005; Wastell, 2001; for an exception see Postert, 2012). But while Luhmann does account for non-modern socialities, most of what he has written about them is quite problematic. Thus, Luhmann paradoxically works best for anthropology if one leaves aside most of what he has to say about segmentary societies, ritual or similarly classical issues.

Therefore, I employ Luhmann eclectically. His highly differentiated and unique use of theoretical terminology and its rigid systematization suggest that it is difficult to extract just some of his terms from his work without having to deal with the whole package. However, as Luhmann has repeatedly stressed, his is a multi-centric theory that does not hierarchize its various terms. This provides me with a certain freedom to employ what is useful while leaving out what seems less so. I will focus on three aspects of Luhmann:

– the separation of psychic and social systems;

I also diverge from him in some respects. Luhmann argues that only consciousness can irritate social systems, that is, make them produce communications. However, he does not assume that consciousness needs to be present on both sides of a communication, thus rejecting phenomenological approaches like those of Alfred Schütz (Luhmann, 2008: 49–50, see Schütz and Luckmann, 1975, ch. II, B, 5, a). Thus far I agree with him. However, he states that present-day society assumes co-presence of consciousness on both sides, while other societies – more simple, or evolutionarily different – do not necessarily make this assumption and therefore allow for communication with gods, animals, ancestors and other non-humans. ⁶ There is, however, no reason to uphold such a difference between ‘modern’ and ‘non-modern’ societies. I hold that the analysis of the inclusion of non-humans that I derive from Luhmann applies in principle to any society. In this respect, I combine Luhmann with the symmetric anthropology of Latour. This informs other divergences from Luhmann, elaborated below.

Psychic and social systems

Luhmann’s theory of communication and autopoietic, that is, self-reproducing social systems, provides a framework to analyze movements between ontological states. First, I will argue that communication is not necessarily bound to human personhood, on the basis of Luhmann’s differentiation of psychic and social systems. Autopoietic systems produce their own elements, but not those of another system. The elements of psychic systems are thoughts – or, more comprehensively, intentional acts (Luhmann, 2008: 31) – while the elements of social systems are communications. These two types of elements (operations) are entirely different from and do not turn into each other. Psychic systems cannot communicate with each other, although they can irritate social systems, spurring them to produce communications.

Thus, psychic systems are in the environment of social systems, and vice versa. As a social system cannot take its elements (communications) from any other system or its own environment, ‘only communication can communicate’ (Luhmann, 2008: 109, my translation). Only select thoughts irritate the social system under specific conditions, and these conditions are set by the social system (Luhmann, 1984: ch. 7, II). This difference between psychic and social systems seems to reproduce the arch-modernist distinction of mind and body, or individual and society. However, I suggest that it can be taken beyond this restriction to modernity.

All communications are by definition elements of a social system. Assuming that no society today is isolated, Luhmann states that the most encompassing social system virtually spans the globe. All other systems are subsystems that emerge due to a number of types of differentiation – segmentation, stratification, center-periphery and functional differentiation being the most prominent ones (Luhmann, 1998: ch. 4). Thus, villages are social systems in segmentary differentiation (1998: 601; see also Sprenger, 2008), but so are large functional fields like religion and politics in modern, functional differentiation. Even temporally and spatially bounded events – a ritual, a gathering, a visit among neighbors – are social systems, although they might be volatile and not fully autopoietic in themselves. These examples are what Luhmann calls interaction systems in which actors are present (Luhmann, 1984: ch. 5, IV) – other than in functional systems, where they are often remotely connected through writing, etc.

The kinds of animation and personhood that are the subject of this article mostly unfold in interaction systems. As I will argue, animation and personhood are processes of making actors present in interaction. This demands a specific semantic, a set of cultural notions about persons, spirits, life, etc. Therefore, these interactions emerge within larger social systems – villages, (vaguely defined) ethnicities, religious communities, etc. Especially in Southeast Asia, the main focus of this article, the techniques of making spirits present provide a transcultural ‘lingua franca of localism’ (O’Connor, 2003: 282). These larger systems provide the conditions for the emergence of those impermanent, local and often ritual systems that actualize non-humans as animate and person-like, each of which is, like all social systems, transient and improvisational (Luhmann, 1984: ch. 1, III).

What, in relationship to such systems, is a subject? Each communication is a selection from a virtually countless number of possibilities, some more, some less probable, conditioned by what has happened before (Luhmann, 1984: ch. 1, II; ch. 4, IX). Thus, each communication creates the conditions for follow-up communications. In order to do so, it needs to be bounded. This closure is what Luhmann calls self-reference. The closure and finality of one communication formally communicates that another communication, a reply, a contradiction, a virtual denial might occur next. Self-reference thus implicitly runs along everything else that is being communicated (Luhmann, 1984: ch. 11, III). Otherwise, there would be no difference between one element and another.

The self-reference of communications in social systems is functionally equivalent to self-awareness in psychic systems. Thus, the most important indicator for subject status is not restricted to psychic systems. Self-awareness is only the specific form that self-reference takes in psychic systems, but subject status exists beyond it, via self-reference. As a consequence, the subject-object divide is suspended. There is neither a necessity to apply it to communication nor any fixed manner to do so (Luhmann, 1984: ch. 12, II). The distinction of subjects and objects, or any other type of being involved in or excluded from communication, thus appears as a result of communication, not as its condition.

As social systems produce only communications, they produce, with each element, the distinction between communication and non-communication. What is not communication is part of the environment, and therefore, while producing communications, social systems produce their own closure as well. Therefore, the self-reference of the elements (communications) – their difference from other elements – is virtually the same as the self-reference of the system – the distinction between elements and non-elements, which defines the boundary between system and environment (Luhmann, 2008: 27). Autopoietic systems thus feature the basic conditions of subjects, making the production of subject status an ongoing effect of communication (Luhmann, 1984: ch. 12, II).

For this reason, Luhmann often speaks of systems and communications in the active voice. Both social systems and psychic systems can function as observers. Observers make the differences that are the base of the selections of follow-up communications, thus enabling their production (Luhmann, 2008: 47). Recognition of communication, or ascription of personhood, is thus not necessarily a cognitive act of consciousness. In a social system, this is a specific relation between one communication and its follow-up communications. Recognition of an event as communication or only as perception shows in the way the event conditions (follow-up) communication. This difference between communication and non-communication is made by an observer – and, according to Luhmann, social systems, as they are able to make the difference, are observers. Thus, one could argue, if only consciousness can make social systems produce communications, as mentioned above, a single, socialized consciousness is sufficient to do so (see below).

At this point, non-human beings enter the picture. A social system is unable to recognize a psychic system by thinking, that is, in terms of the psychic system’s operations. When social systems recognize psychic systems in their environment, they can only do so in terms of communication. As communication is semantically coded, ‘psychic system’ is only one culturally specific concept for this entity or system in the social system’s environment – other terms might be ‘consciousness’, ‘subjective agent’, ‘person’ and a multitude of others if we would step outside English academic discourse. If I keep the terms ‘psychic system’ or ‘consciousness’, it is for the coherence of the argument as part of this discourse.

But a social system cannot ‘see’ the operations of psychic systems in its environment, only recognize their impact on its own reproduction of communication. Their opacity is a condition for the social system to reproduce. Social systems always operate ‘as if’ there are psychic systems in their environment (Luhmann, 2008: 33). This point offers another reading of Luhmann’s claim that only consciousness, i.e. psychic systems, can make social systems produce communication, a reading more radical than my statement above that a single consciousness is sufficient for communication. The indicators of the presence of consciousness, or psychic systems, or subjective agents like persons, are specific to the semantic of the social system and vary accordingly (Luhmann, 1998: 643). This allows for social systems in which notions like ‘consciousness’, ‘psyche’ or even ‘body’ are not the most important conditions for communication to take place. In which sense, then, could ‘consciousness’ still be a requirement for social systems, if it primarily appears in their self-description?

In any case, the required subject status, resulting from self-referentiality, is a constitutive aspect of social systems themselves. Therefore, non-human beings cannot be by definition excluded from social systems (Luhmann, 2008: 51). The status of subject in social systems is not fixed to a specific type of agent, like living human beings. Rather, it is dispersed across reality, as far as communication is concerned. However, social systems consist of communications, not persons. How, then, is personhood created out of communications?

Communication and personhood

Before elaborating upon this point, I need to detail Luhmann’s definition of communication. Communication consists of three symmetrically arranged, non-hierarchical aspects (‘selections’). These are information, utterance, and understanding/misunderstanding. Information refers to the content of the communication, utterance to its form and context, and understanding/misunderstanding to its reception (see also Sprenger, 2011). ⁷ Social systems distinguish communication from mere perception by making a difference between information and utterance – a follow-up communication is directed towards an uttering entity, not to the information (Luhmann, 2008: 49). But it is reception (understanding/misunderstanding) that completes the communication and makes it connective to follow-up communications. For this reason, Luhmann chooses to call the recipient Ego and the sender Alter (Luhmann, 1984: ch. 4, II).

This is crucial for communication with non-humans. One is reminded of the anecdote about thunderbirds that one of Irving Hallowell’s Ojibwa interlocutors related. During a severe thunderstorm, an old man turned to his wife and asked: ‘Did you hear what was said?’, to which she replied: ‘No, I didn’t catch it’ (Hallowell, 1960: 34). Here, it is the listeners who acknowledge the sound of thunder as a communication by the thunderbird. They do so independently of whether they understood it or not. The dialogue of the Ojibwa couple reveals that it is not decisive to grasp the information of a communication in order to make the distinction between the utterance – which they recognized – and the information – which they did not get. The identification of illness as communication by spirits in Southeast Asia is another example of the same process (see below).

This leads to the next step of my argument. The terms Ego and Alter, for Luhmann, do not yet specify persons or human beings. They simply denote the horizons implied in each communication, regarding sender and recipient. They leave open the question; do these two parties consist of human beings, social systems, organizations or other kinds of entities (Luhmann, 1984: ch. 3, I)? ⁸

However, each social system needs to do the work of attribution. It has to identify Ego and Alter in order to secure the reproduction of its elements. Thereby, it asymmetrizes the symmetry of information, utterance and understanding, by giving communication a direction – from Alter to Ego. In other words, in order to reproduce, social systems have to describe their operations as actions performed by agents. Only in this way, by replacing the symmetry of the three aspects with the asymmetry of sender and receiver, can one communication lead to another (Luhmann, 1984: ch. 4, VIII).

This fills the empty slots of Ego and Alter with definitive entities and thus creates actors. These do not need to be concrete human beings. As soon as an observer – social or psychic system – differentiates information and utterance, a social system produces an Alter as a kind of retrospective cause of the communication. Thus, paradoxically, what appears as the cause of the communication for the observer is in fact an effect. Observers situate agency with some bundle of conditions and attributions that emerges to constitute a sender because it produces something that is recognized as communication. Something comparable occurs when human actors blame their actions or inaction on ‘the system’ which conditions their respective communications or communicates itself, e.g. bureaucracy – note that such ‘systems’ are part of a vernacular and have only fleeting similarity with Luhmann’s systems (see Gershon, 2006).

On the level of the social system, the attribution of communication to senders and receivers – that is, the assumption that communication is action – is not a cognitive act of individual actors. Rather, communications in themselves assume the existence of communicating agents. This way, a series of communications agglomerate to create the shape of actors. Senders and receivers are thus somehow bounded entities that come about as communications-as-actions seek out beings to be attached to. The internal self-references of communications constitute these beings as subjective agents. Persons thus are ‘highly aggregated self-references’ (Luhmann, 1984: 182; ch. 3, IX).

The form of personhood is, according to Luhmann, particularly favorable as a way to attribute actions. Each communication is highly specific and implies a virtually endless variety of follow-up communications, but in most cases this leads to fairly predictable results. A social system accomplishes this by reducing the complexity of communication by introducing certain conditions for its autopoiesis. This favors forms of attribution with a high potential to actually process complexity and stabilize the results, but at the same time function as condensations filling the slots of Ego and Alter. These forms of attribution state the existence of particular types of beings whose behavior is at once to a degree predictable but also open to variation.

Personhood is such a form, or rather, a category of such forms. As the semantics of social systems often indicate personhood with signs of agency, intention, accountability and other socially relevant features, personhood is able to account for a great variety of contingent communications. Social systems thus can attribute personhood to beings which are communicative others in interactive communications or in suspended ones (e.g. writing). It even allows recognizing those others as beings that either reveal their intentions or hide them (see Luhmann, 1984: ch. 3, II).

Therefore, systems can register either empathy/transparency or opacity between Alters and Egos, which in turn is an important means to attribute a great range of communications to persons. Researchers have observed the importance of ideas of opacity in such different contexts as Melanesia (Robbins and Rumsey, 2008), Thailand (Moerman, 1969: 457), Mesoamerica (Groark, 2013), and techno-social relationships (Rammert and Schulz-Schaefer, 2002: 18). Also, the Rmeet uplanders I know often stress their lack of understanding of the spirits they address in their rituals. This is comprehensible when opacity and empathy are understood as a difference that is crucial for the attribution of communications. Registering empathy stresses the openness of persons and the similarity among communicators, while opacity emphasizes their difference and unpredictability. Both are ways to differentiate utterance and information and thereby enable the recognition of communication. In their combination, ideas of empathy and opacity make it possible to define persons as the sources of unpredictable, even incomprehensible communications – and thereby widen the range of events that an observer can recognize as communication.

By employing the difference of opacity and empathy, social systems acknowledge that they are unable to process the complexity of a person in its entirety as communication – by definition, a psychic system, consciousness, etc., is external to the social system and therefore always out of reach. Personhood and its aspects thus refer to the complexity of a being that irritates the social system from its environment. Opacity and empathy signal opposite tendencies of access to it, or rather, opposite possibilities to process the complexity of the (assumed) other system by the social system. Persons thus ‘condense as a side effect’ of the operation of social systems (Luhmann, 2008: 143, my translation).

The way Egos and Alters come into being is, however, dependent on the particular semantics by which social systems reproduce communication, that is, on culture (Luhmann, 1984: ch. 4, VII). The form ‘person’ emerges differently under different semantic conditions, as social systems recognize and attribute communications in different ways, and it is not the only form that responses to action/communication can take. Different kinds of beings might condense along the process of communicating and translating communication into the form of action. This ultimately brings us back to the initial question. How are beings constituted as animate or person-like – and how do we account for cases in which they fluctuate between personhood and mere animation?

I hold that situationally different recognitions of communication are involved. It is the contingencies of the interaction with non-human beings which create shifts in their ontological or epistemological status (see Bird-David, 1999). As the relationships change, the recognition of events as communications and their attribution varies. Recognition and attribution are indicated by the kind of communications that the respective social system produces in response to certain events in its environment. Thus, communications sometimes identify beings as if they can be manipulated in an almost mechanical manner, e.g. by directing the flow of life forces through ritual (communication about). At other times, the communications directly address them, as the recipients of propitiating gifts and pleas (communication with). In the first case, the system produces follow-up communication as if no opacity or hidden intentions existed in the non-human. In the second, it differentiates utterance and information in any event connected with the non-human (see Luhmann, 1987: 228).

This is not a contradiction. Rather, it indicates a difference in the conditioning of communications. Observers (social or psychic systems) attribute events to these beings that they sometimes treat as communications and in other contexts as effects of processes that are not communicative. Recognition, however, is always communicative and thus social (see also Fuller, 1994: 748). This means that shifts on the PAI axis between object, potency and person do not need to be huge, category-defying leaps, e.g. from mere animation to personhood and back. Of course, sometimes such movements cross the boundaries between semantically encoded categories of beings. But the movement might also occur in more subtle steps. Each single communication can contribute to the emergence of animation or personhood or lead away from it. Within the range of communicability, persons can be differentiated according to their defining features – some can be addressed in vernacular language, others only in ritual, some in waking life, others in dreams only, some being predictable and accountable, others not so. They are thus persons on a qualitative, graded scale of personhood (Sprenger, 2016b). Because such varieties are closely bound to specific semantics, I will continue my argument in the form of an ethnographic example that shows how the emergence of different kinds of beings is conventionalized in ritual procedure.

Rmeet agricultural spirits

The Rmeet (Lamet, Rmet) are Mon-Khmer speakers who mostly live as swidden farmers in upland Laos. They belong to the more than 30 percent of non-Buddhists in Laos that practice localized ritual systems. In these rituals, they address in particular ancestral and territorial spirits through gifts, usually of food, animal blood and liquor. ⁹ Agriculture combines the physical interaction with the land, plants and animals and the ritual interaction with spirits. The latter always precedes interaction with other non-humans at each stage of agriculture. Communication with spirits is thus a necessary condition for the tilling of the land and the growth of crops (Sprenger, 2006, 2016c). Divination and offerings are the most common techniques to establish any kind of spirit as a communicative agent, and Rmeet indeed call offerings and ritual exchanges ‘making the spirit’ (plo phi). Outside of rituals, spirits have a vague and somewhat undefined existence. They sometimes consist of bundles of relationships. A single house spirit, for example, is the agglomeration of male lineal ancestors with their wives, while the village spirit is the agglomeration of all house spirits. Rituals that ‘make’ such spirits are thus examples for the processes that create non-humans through communication.

Two types of spirits are of crucial importance: the male spirits of the earth (jom or phi mäa) and the female spirit of rice (phi ngo). These two differ in the way the Rmeet communicate with them, which is to say that they are different kinds of beings. In many ways they are contrasting. The semantic differences that position them on a specific axis towards personhood are: anthropomorphic/non-anthropomorphic, localized/diffuse, intrusive/non-intrusive.

Spirits of the earth, the ‘lords of soil and rock’, are dangerous and only good as long as they leave people alone. They have an almost person-like definition, with the potential of having distinct bodies. They might appear in dreams, after a farmer has decided to clear a particular plot, in the shape of soldiers. Like soldiers shooting projectiles into bodies, earth spirits cause illnesses that shamans suck out of their patient’s body in the shape of small stones or pieces of wood (Sprenger, 2010).

On important occasions when earth spirits receive gifts, they are embodied as clay figures, marked by their rounded heads. A ritual performer, e.g. a member of the household tilling a field, sculpts them on a tray that also contains offerings – small pieces of tobacco, liquor, fermented tea for chewing, steamed rice. In the course of the ritual, the figures receive the fresh blood of a chicken that a household member of the sick person kills and cooks. Ritual performers first serve meals to the spirits, while humans might participate afterwards. Earth spirits receive these offerings either as a cure for an illness diagnosed by a shaman (or diviner) or at particular stages of the agricultural cycle, in particular before harvesting. The subsequent disappearance of illness or its absence then counts as responsive communication – the spirit has accepted the reciprocity implied in the gifts.

Rituals are interaction systems that establish the communicative parties as Egos and Alters. Illness in any case spurs the social system into producing communications. These consist of divination or shamanic sessions. Shamans are capable of going into trance and then seeing or hearing the various spirits. Their first step is to identify an illness as communication or non-communication, as there are types of illness not involving spirits and without a clear etiology. The shaman does so by employing various types of divination or by going into trance to communicate with the spirits directly. If his enquiries with various spirits remain inconclusive, he does not recognize the illness as communication. Here again Ego, the receiver – diviner or shaman – determines if an event is communication. If this is the case, he has to differentiate between utterance (illness) and information (the spirit’s demands). To this end, the shaman specifies the communicating agent as either a spirit of the earth, of the sky, a dead person, etc. He then goes on to inquire about the spirit’s demands and to negotiate the human response. For this latter step, going into trance is almost inevitable, and this usually involves a larger gathering at the sick person’s house. In this process, as more people get involved and the density of communication increases, the spirit acquires an increasingly differentiated shape – until performers provide it with a physical body, a clay figure. Embodying the spirit is thus in itself a communicative act that enhances the spirit’s responsiveness and personhood. The shaman wipes the sick person’s body and invites the spirit to move into the clay figure. Relatives of the sick person then bring the spirit to his place in the fields or forest, where he shares food with the human beings. The spirit becomes a temporary member of the village, marked by commensality, as a guest who, after the meal, is supposed to return to his own village.

Communication with the spirit thus involves the application of a series of semantically specific distinctions – communication and non-communication, types of spirits, of demands, of human responses. Thus, each communicative act further condenses the spirit, be it as sender or as receiver, as each communication creates constraints and potentials for further communications. The communications with the spirit accumulate self-reference that takes the shape of the person of the spirit. The spirit is communicated into being a person, as it were.

This also applies to a second effect. The accumulation of communications increasingly draws the spirit into human society. There is an important differentiation that the Rmeet observe between beings belonging to their own and to different communities. Spirits, in particular those of the dead and the land, have their own villages and relationships. This differentiation parallels and amplifies that between other communities in the vicinity, ranging from culturally somewhat different Rmeet villages to those of other uplanders to lowland towns. Such differences among villages and communities mark differentiated potentials to communicate (see also McKinley, 1976; Platenkamp, 2007; Sillander, 2016). Spirits are thus non-social only in relationship to human communities, but maintain their own social systems, which are mostly inaccessible to humans. Therefore, human social systems like villages and other communities need certain institutions that address and process this difference between systems, like divination, shamanism and sacrificial rituals. Communicating with a spirit draws him into the horizon of the system of the respective village or household, if only temporarily. ¹⁰

The rituals performed before harvesting follow procedures comparable to those performed for an illness. The male head of the household tilling the field or an agnatic relative places a tray with round-headed clay figures into the field and sacrifices a chicken. This way he asks permission to use the field and begs the spirits not to harm people. As in the healing rituals, the relationship of the earth spirits to a particular place is recognized. Potential personhood and locality thus are linked, both being modes of differentiation.

Matters are different with the rice spirit. Unlike many societies in Southeast Asia, the Rmeet do not seem to tell myths about the spirit of rice, for example as a rice goddess (e.g. Terwiel, 1994). While the rice spirit (phi ngo) is female, she is a rather diffuse entity incapable of actively harming people, except by refusing to join the rice on its way to the granaries. When the rice spirit is absent, the rice ears will carry empty husks, or the rice stock will deplete quickly. Although the household tilling the field secures the presence of the spirit by ritual offerings, she is in between personhood and potency.

The rice spirit differs from earth spirits in other respects as well. Besides not communicating on her own initiative, e.g. by causing illness or appearing in dreams or trances, she is also not embodied by clay figures. While earth spirits belong to particular places and plots, the rice spirit needs to be gathered together. This is the task of women, in contrast to the men, who address the earth spirits.

On the first day of harvesting, the mother of the household that is tilling the field, who is now called ‘rice mother’ (nee ngo), gathers with a number of women somewhere at the center of the field. She places a tray on the ground containing rice, tobacco, chewing tea, and (in some villages) the blood of a chicken. At the same time, she reveals the household’s collection of ‘rice medicine’ (chenüm ngo) to the spirit, thereby attracting her to humans. ‘Rice medicine’ mostly consists of heirlooms and odd objects that are marked by their ancient, unique or unusual nature, like cobblestones found far away from water, seashells, Buddhist amulets, and the teeth of ancestors. These are not bodies for the rice spirit in the same way the clay figures are for the earth spirits. However, they support the localization of the spirit by drawing her to the location of the ‘rice medicine’. Also, the keeping of durable heirlooms embodies the continuity of the relationship with the rice spirits, while the clay bodies of the earth spirits are impermanent, just like their relationships with humans are supposed to be.

For this reason, the items are usually stored in the granaries and only taken out for the harvest. At this time, the ‘mother of rice’ calls the rice spirit from faraway fields, enumerating all the lands of the neighboring villages. The rice spirit is diffusely distributed across the landscape and the ritual serves to concentrate her. Tillers channel her movements on the field by logs across footpaths smeared with chicken blood. At the end of each harvest day, the house mother or father ties a bunch of standing rice ears in a knot and venerates it, thus temporarily enhancing the concentration process towards a non-anthropomorphic body. After the harvest, they place a bundle of ears containing the spirit into their house or granary. ¹¹ While the localization of the earth spirits forms the base of their personhood, the localization of the rice spirit is subject to a constant communicative effort. The success of these delicate measures only shows in the course of the year, depending on the depletion of the stock. Still, this spirit does not attain personhood to the degree that the earth spirits do. For example, I received contradictory answers to the question: is the rice spirit is one or many? – both answers seemed plausible. She is thus closer to the kind of impersonal life force or potency that is subject to manipulation in many other Southeast Asian rituals. These forces are, while elusive, usually conceived to be open to direct manipulation.

The earth spirits and the rice spirit among the Rmeet are thus situated towards the more person-like end of the local axis of beings. Earth and rice spirits occupy different positions on it, but in the course of the ritual exchanges move within a specific range. There is no marked threshold that separates the modes of communication for the two spirits into different types, as some of the rituals used to address them are similar. This means that their movements along the axis overlap. But their difference appears as a series of semantic distinctions that are sequentially realized: localization vs. diffusion, illness-inducing (i.e. communicating on its own initiative) vs. harmless, anthropomorphic bodies vs. non-anthropomorphic bodies. Their ritual treatment shows how the Rmeet communicate spirits into being by gradually socializing them into the village. However, this is only possible because spirits are, while mostly invisible, still treated as communicative and thus part of an encompassing social system. This system contains the villages of the living as well as the spirits, even though boundaries and types of communication differentiate it internally.

Persons and potencies

This analysis speaks against assumptions that personification occurs because what is being personified is either important (Tiele, 1897), unusual (Bird-David, 1999; especially Viveiros de Castro’s comment on p. S80) or unknown (Guthrie, 2013; Marett, 1914: 12). Even Luhmann, being the modernist thinker he is, shares this assumption when he writes of the human tendency ‘to reduce everything that is unknown to the model of “persons”’ (Luhmann, 1984: ch. 5, IV, my translation; see also Luhmann, 1995: 188). From the perspective of the current analysis, there is no causality in this. Social systems identify importance or unusualness just as they identify communication. The way they recognize communication, however, has nothing to do formally with the way they identify importance. The question is, rather, if communication is about these beings or with them (Luhmann, 1987). This brings us back to the initial model. The ethnographic example served to demonstrate how personhood can be graded and how this graded existence comes into being. Thus, animist forms of communication do not treat non-humans constantly as persons. They might shift between personification and objectification as well. As Remme (2016) has argued, this is indeed a constitutive aspect of animism (see also Karim, 2004). Therefore, attribution of life as a form of social presence varies according to the communicative situation. The question remains how beings beyond the horizon of personhood, those that are merely animate or inanimate, gain their respective ontological status. This also concerns forms of potency that are – often ritually – manipulated.

Communication with such beings automatically establishes them as Egos/Alters. If communications are restricted to those about them, social systems relegate these beings to their environment. However, social systems process events in the environment as information depending on the sensitivity or resonance of the system, that is, depending on their semantic (Luhmann, 1990: 42–3). The distinction between ‘communication with’ and ‘communication about’ seems to demarcate social systems and their environment quite clearly. However, the distinction between persons and objects is the result of a series of much finer-graded communications employing additional distinctions, similar to the difference between earth and rice spirit. I thus do not argue that the ‘line’ between person and object is ‘blurred’. Each communication employs specific distinctions that are, in themselves, clear enough. However, as each communication is situational, it denotes the ontological status of beings regarding personhood precisely, but only temporarily. It is conventionalization and the repetition of the same set of differences that stabilize the ontological status (Luhmann, 2008: 33).

However, as I have argued, ongoing communication does not always employ the same set of differences. The impression of ‘blurring’ – which scholars looking at local communication might have when they expect ontological fixity – would result from an attempt to superimpose different moments of existence, that is, different communications, onto an ontologically fixed shape. As the delineation of a being changes in the course of communication, the differences employed possibly do not match. The attempt to analytically fixate ontological status thus misses the nature of communicating beings, which are, like their communications, temporalized and transient.

Although the distinctions are part of culturally specific semantics, I suggest a few generalizations. The most important of the distinctions made in the course of establishing the status of beings is communication/non-communication. The self-reference of communication then is able to elicit follow-up communications. In the case of spirits, such follow-up communications might take the form of verbal utterances, but also of gifts and offerings. Yet, ritual communication (with) and ritual manipulation (of) are very close, and therefore, non-human persons and potencies are just one communicative step apart. Social systems thus often do not fix personhood in advance but offer the opportunity to assign it through cumulative attributions of communications. In particular, ‘animist’ personhood is an open, contingent project.

Thus, there are two tendencies in the communication chains. One consists of certain expectations regarding the form that the chain will ultimately take, e.g. it might proceed toward an outcome like personhood. Such expectations might impose a certain recursive teleology upon the chain (Luhmann, 1984: ch. 8, XVII), as the efforts of Rmeet shamans to identify a spirit as the cause of disease suggest. However, each single step is a selection from a great variety of possibilities. As the reproduction of social systems is very much a matter of contingency and improvisation (Luhmann, 1984: ch. 3, II), they could select different communications any time. The process of attributing communications to persons and recognizing non-humans as actors might break off at any moment. This is the other tendency. Both tendencies together produce what I have called graded persons. These are variously alive and personified, according to the contingencies that shape communications.

This draws attention to the steps in between mere objects and communicating beings in the scheme I propose. Within this range, another important difference is negotiated through communication, the differentiation of life from dead matter. An empirical demonstration of how this differentiation is accomplished would demand another article. Here, I want to make just a few points. First, the general assumption that everything is alive, as brilliantly explored by Ingold (2006), is philosophically attractive but analytically unsatisfactory, as it produces a kind of hyper-complexity. If everything were equally alive, all relations between beings would be potentially equal, and everything would be connected to everything else. This is quite evidently not the case. Any kind of system – not just social ones – is based on differentiation (i.e. production of complexity), reduction of complexity and stabilization (see also Bovensiepen, 2014). ¹²

This means that the recognition of life demands that the social system operates with the difference of living/inanimate. This difference can be situated along an axis like the one I have sketched out above. Living beings appear different in social systems than inanimate ones. They are also closer to becoming persons than inanimate ones – and this includes spirits of the dead, insofar they are not inert but still responsive. ‘Life’, however locally conceived, is thus another approximation to sociality.

One possible axis allows beings that are ‘alive’ to move towards beings that possess life-force, or potency that can be localized and manipulated, or even towards personhood. Life-force and potency are thus forms of life which are subject to particular, additional distinctions. Life and potency are separated by small, specific steps, in which distinctions like localized/diffuse are employed. This way, beings like the Rmeet rice spirit condense between potency and person.

Concluding remarks

The aims of this article have been to understand how beings are constituted as communicative and agentive and to explore the potentials of Luhmann for the current debate about the sociality of non-humans. It suggests an alternative to approaches like those of Bruno Latour or Tim Ingold. The latter have significantly widened the range of relevance of anthropology, but do so by focusing on sometimes ill-defined processual flows and a permanent merging, almost denial, of conceptual boundaries. Thereby, they tend to miss or at least under-appreciate the productive capacity of differentiation – even though differences are all we can immediately observe. They also tend to fix the ontological status of beings, disregarding observations that beings are sometimes alive and agentive and sometimes not, sometimes communicative and person-like and sometimes not.

The approach I suggest here does not fall behind Ingold and Latour, for example by claiming any privilege to Homo sapiens. Instead, it addresses the question of how non-humans become agentive and person-like in collectives by focusing on communication. It specifically highlights beings which, other than the sensuously present animals or technological devices that Latour and Ingold tend to speak about, are invisible and often distant – souls, spirits, gods. To understand the way these beings – and by extension, persons in general – come into existence, I argue, it is necessary to reverse the conventional relationship between actor and communication.

Along with Luhmann, I distinguish communication from thought and locate self-reference and therefore subject status in communications. The recognition of events as communication emerges from the follow-up communications that they provoke. Thus, subject status and communication are not primarily a matter of ‘consciousness’ or ‘cognition’, but of the reproduction of communications as elements of social systems. However, any communication needs to be ascribed to a person in order to condition its successive communications. Therefore, persons emerge as ‘side effects’ of ongoing communication. Life and personhood result from the way social systems treat events as communication with or about beings.

But while Luhmann’s distinctions between psychic and social systems seem rather anthropocentric and all too familiar to those who are ‘modern’ in Latour’s sense, this article attempts to break some of Luhmann’s concepts away from modernity. The non-humans excluded from participation in society by modernity for lack of discernible psychic systems can be reintegrated by the very means of their exclusion, the presence of communication. By focusing on communication, Luhmann has called the distinction of mind and society into question. If only communication communicates and comes to completion only by reception, the necessary relation between consciousness and communication, stated by Luhmann, appears as an assumption that certain social systems make as a function of their culturally specific semantics. Notions like ‘consciousness’, ‘thought’, etc., appear as devices that are necessary for the ascription of agency and personhood but that are culturally variable.

This radical reading of Luhmann suggests a gradual emergence of non-humans as social beings and persons that unfolds through a process of self-referential reproduction of communications and their agglomeration. The recognition of opacity in others, the employment of terms that denote aspects of personhood and other communications add up to make persons – they are communicated into being. Thereby, communications with non-humans, in particular invisibles like spirits, lose the illusory character that has often been ascribed to them. This implies a radical shift away from individual intentionality as the constitutive force of sociality. It allows beings, person-like or alive, to shift between ontological statuses. This is, however, not just a consequence of theoretical ruminations, but firmly established in the ethnographic record, e.g. on Southeast Asia.

This in turn implies some avenues for further research. Models analogous to the PAI axis – that may be axial, multi-axial, concentric, flow-charts, etc., with different terms and boundaries between them – could serve as the basis of specific analyses. The analytical procedure, however, would be the same. Such models account for various extensions and contractions of the notion of personhood and life. While I have mostly focused on animist relationships and spirits, I believe that the same approach will yield results for other kinds of non-humans and humans as well, thus enabling comparison beyond fixed typologies and schemes. For example, this kind of analysis might account better for the co-presence, in Western modernity, of naturalist and ‘animist’ relationships. While modern biology, as Ingold has noted, has a far more restricted notion of life than animist cosmologies (Ingold, 1990, 2006), this does not apply to all of modern society where – sometimes playful – attributions of life to cars, computers or stuffed animals occur.

Heeding the movements along these models implies asking a number of questions in any given ethnographic situation. First, what are the concepts of person and their grades of animation and inanimateness? Second, which specific distinctions are employed in which way to make a being shift along the axes, and in which order? Third, for how long and under which circumstances are the respective states stabilized? As communications are impermanent, the ontological status of a being needs constant reinforcement – or it will change.