What is Spirit Possession? Defining, Comparing, and Explaining Two Possession Forms

Click to join the conversation with over 500,000 Pentecostal believers and scholars

Click to get our FREE MOBILE APP and stay connected

ethnos, vol. 73:1, march 2008 (pp. xx–xx)
What is Spirit Possession?
Defi ning, Comparing, and Explaining
Two Possession Forms
Emma Cohen
University of Oxford, UK
abstract Reviewing anthropological analyses of possession forms cross-culturally
and drawing from recent advances in cognitive psychology, this paper attempts to
explain recurrent features of spirit possession. Spirit possession concepts fall into
broadly two varieties: one that entails the transformation or replacement of iden-
tity (executive possession) and one that envisages possessing spirits as (the cause
of) illness and misfortune (pathogenic possession). The cross-culturally recurrent
features of these divergent conceptual structures may be explained, at least in part,
with reference to distinct processes of human psychology, one set of which deals
with the representation of person-identity and another that deals with notions
about contamination.
keywords Comparative anthropology, person-identity, spirit illness
In
1976, Erika Bourguignon, one of the foremost anthropological scholars
of spirit possession, published a book entitled Possession in which she
presented a cross-cultural analysis of possession beliefs and behaviours.
Bourguignon parsed the broad range of possession phenomena into two types.
Both types entail the belief that a ‘person is changed in some way through
the presence in him or on him of a spirit entity or power, other than his own
personality, soul, self or the like’ (1976 : 8). The two types were labelled ‘pos-
session trance’ and ‘possession’. Possession trance is expressed in altered states
Possession trance
Possession trance
of consciousness; in possession, such trance states are absent. This particular
method of carving up the domain proved useful for assessing the frequency
and distribution of possession, possession trance, and trance (an ‘altered state
of consciousness’ entailing no associated possession belief ) in cross-cultural
survey of a world-wide sample of 488 societies (Bourguignon 1968).
2
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
Such apparently tidy descriptions were widely criticized, however, for their
failure to capture the complexity, variability, and polysemy that characterise
actual representations of possession on the ground. Possession forms, like
many other topics of anthropological study, resisted being pinned down
to a singular, one-size-fi ts-all defi nition. Indeed, it was stressed, abstract
context-free defi nitions of apparently widespread — but different — cultural
phenomena, by their very nature, gloss over the very specifi c, unique, cul-
turally-embedded qualities of possession phenomena in their local contexts.
And these qualities, it was proposed, were what anthropological enquiry
should be concerned with. As historical and cross-cultural comparative and
explanatory approaches increasingly gave way to particularistic interpretive
accounts in anthropology, the relevance of a generalizable defi nition of what
counts and what does not count as possession declined. The central conten-
tion was that possession was no longer considered a ‘thing’ to be defi ned or
prised apart and dissected from the ‘whole’ within which it could be more
adequately and faithfully understood and represented – the parcelling of
the socio-cultural domain according to arbitrarily selected and culturally
insensitive measures, labels and divisions risked compromising the fi delity
of the holistic interpretation.
The general discussion is a familiar one to anthropologists across the
discipline, capturing the essence of crucial concerns about the tools of
description, comparison, and generalization in the generation of anthropo-
logical theory. An enduring and valuable development in anthropological
scholarship ever since has been to focus on the interpretive understanding of
the constitution of dynamic, symbolic, socio-cultural worlds and the beliefs,
concepts, ideologies, institutions, confl icts, and so on that compose them — a
broad range of approaches referred to ‘critical hemeneutics’ (Lambek 2002:
6). However, insofar as these anthropological studies of particular cases of
particular people in particular places and contexts continue to endeavour
to speak to one another, even at a basic descriptive level, the thorny issues
of defi nition necessarily remain uncomfortably salient. What, if anything,
unifi es what appear to be recognizably cross-culturally, recurrent features
of our pseudo-analytic concept of ‘possession’? What, if any, are the criteria
by which one might assess the comparative utility of a study of ‘possession’
in one part of South East Asia for the development of an understanding of
possession phenomena in another part of the region? How might one expli-
citly set out the dimensions along which similarity and difference may be
reliably measured? Can common features be objectively identifi ed or were
What is Spirit Possession? 3
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
Bourguignon’s critics correct — any such typology leads to distortion via the
imposition of inappropriate categories?
This paper offers a fresh approach to these core issues and concerns. Any
comparative, theoretical approach requires a well-circumscribed description
of what phenomena may be usefully and legitimately compared, and of
what the theory applies to (and what it does not). In this sense, (tentative)
theory informs (the working) defi nition. Going somewhat against the grain
of interpretivism, this paper presents a working defi nition (or defi nitions)
of possession that is driven by causally signifi cant criteria. It argues that (1)
possession concepts demonstrate cross-culturally recurrent features that
(2) are the product of the mechanisms and processes of regular cognitive
architecture, and that (3) cognitive processes constrain, and therefore ex-
plain in part, the form and spread of these features. In short, notions of what
constitutes possession and the paths by which possession concepts and
practices are transmitted, even across vastly different cultural environments
and historical periods, are informed and constrained by recurrent features
of cognition that guide perception, representation, thought and action (see
also Boyer 2001; Barrett 2004).
This novel but tentative explanatory approach considers possible cogni-
tive causal mechanisms that underpin two core types of possession concepts
that have been widely described in the ethnographic literature. Although
these types track closely on those identifi ed by Bourguignon – possession
and possession trance – they are distinguished according to a different set of
premises. I shall call these two types eeexxxeeecccuuutttiiivvveee possession and pppaaattthhhooogggeeennniiiccc pos-
session. Both possession forms entail the direct actions of spirit entities in or
on a person’s body. PPPaaattthhhooogggeeennniiiccc possession concepts result from the operation
of cognitive tools that deal with the representation of contamination (both
positive and negative); the presence of the spirit entity is typically (but not
always) manifested in the form of illness. EEExxxeeecccuuutttiiivvveee possession concepts mo-
bilise cognitive tools that deal with the world of intentional agents; the spirit
entity is typically represented as taking over the host’s executive control, or
replacing the host’s ‘mind’ (or intentional agency), thus assuming control of
bodily behaviours. The following sections develop an account of the causal
signifi cance of this typology in terms of normal human cognition. Such a
reconceptualization provides a basis for the development of the compara-
tive analysis of like forms of possession cross-culturally and, ultimately, for a
return to explanatory accounts of the emergence, transmission and persistence
of recurrent features of spirit possession concepts and practices.
4
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
Explaining Possession Forms
The fi eld of possession studies is no stranger to analytical typologies.
Possession forms have frequently been defi ned in the context of attempts
to resolve controversies surrounding how possession behaviour might be
adequately interpreted. The distinction between ‘madmen and mystics’ was
a widespread preoccupation in early anthropological and medical analyses
(see Klass 2003 for a recent interdisciplinary approach to these issues). Spirit
possession forms have also been variously defi ned, and typologized, in terms
of their social and communicative functions (Lewis 1971; Firth 1967) and
their structural-functional properties (e.g. Rex L. Jones’ [1976] analysis of
spirit possession in Nepal).
A more neglected issue concerns the recurrences and variations in the
forms that possession ideas take cross-culturally. Bourguignon distinguished
between two forms of possession: ‘one form of possession causes a change in
bodily functioning; the other form of possession alters consciousness, aware-
ness, the personality or will of the individual’ (1976 : 3). Further refi nement of
Bourguignon’s characterization of these forms developed into a distinction
that turned on the presence or absence of trance behaviour. Bourguignon
claimed this to be an important distinction that recurs frequently in the an-
thropological literature. Critics, however, objected that the belief-behaviour
distinction (as it became characterized) was arbitrarily drawn and led to
problematic explanatory analyses premised on a reifi cation of ‘trance’ (the
‘behaviour’ element) and ‘possession’ (the ‘belief’ element) as naturally dis-
tinct. The ill-fi tting categories, it was argued, are then ‘imposed’ upon widely
diverse phenomena in an effort to compare, understand and explain them in
terms of arbitrarily selected principles (e.g. Lambek 1989; Maurizio 1995).
In this paper, I agree that the trance-focussed distinction is an unhelpful
way of analysing the cross-cultural patterns in possession forms identifi ed. The
objections to Bourguignon’s overall approach are not entirely unproblematic,
however. Bourguignon’s ambition to identify and apply comparatively useful
categories that can cut through the diversity of the phenomenon to identify
recurrent, underlying principles and features is potentially of considerable
theoretical value. Indeed, Bourguignon picks out what I will argue aaarrreee im-
portant patterns in the ways in which possession is represented. I contend,
however, that the presence or absence of trance is not a theoretically driven
distinction. Consider the following examples of possession belief, used by
Bourguignon to develop the possession/possession trance typology. Among
the Jews at the time of Christ, beliefs in possession by ‘unclean spirits’ took at
What is Spirit Possession? 5
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
least two principal forms. On the one hand, spirits would inhabit a person’s
body, speaking through the person to reveal their identity. On the other hand,
there were beliefs that entailed the causal attribution of illness or malady
(e.g. deafness, paralysis, etc.) to the presence of the possessing spirit in the
person’s body. A superfi cial analysis of the component features of these two
descriptions of possession could point to any one of a considerable number
of differences, including the presence or absence of trance. How, then, can
we carve the phenomenon at its joints? What should inform our theories
of cultural phenomena? Below I outline the general approach framing the
problems and claims in the remainder of this paper.
Possession is not a ‘thing’ to be explained, but a complex series of patterns
of thinking and behaviour. Failure to recognise this fundamental premise has
led, in part, to the generation of defi nitions (or theories) of possession as
somettthhhiiinnnggg else that purportedly bears arbitrarily selected similarities, or causal
underpinnings (e.g. dissociative identity disorder, or hysteria, or mental illness,
and so on). Possession was considered by proponents of such approaches
to be explainable, for example, with reference to the biological capacities of
humans to experience trance states, or to the largely unknown mechanisms
underlying mental pathologies. It was in response to such culture-blind, magic-
bullet approaches to understanding possession that anthropologist Michael
Lambek presented some alternative ways of conceptualizing possession and
trance phenomena (1989). Lambek’s key suggestion was that not only is pos-
session a whole social complex, but trance also (generally considered as a
psychobiological capacity) is cultural; ‘the appearance of trance is mediated
by the cultural model, by its social reality; the collective representations of
trance precede its incidence’ (1989 : 38). Lambek further suggested that the
variability and complexity of trance and spirit possession phenomena should
be given central place in cross-cultural investigations – ‘in cultural matters,
the lowest common denominator cannot tell us very much’ (ibid.: 37) – and
recommended interpretive over explanatory ambitions. ‘Possession’, he, and
many others since, argued, ‘can enter into virtually all areas of life . . . It can-
not be explained in simple terms. In fact, its very penetration into so many
areas of life, the diversity of its functions and expressions, suggests turning
away from causal, etiological explanations toward examining its structure,
organization, reproduction, and meaning’ (ibid.: 45). I suggest that although
the turn toward context–sensitive interpretive accounts has richly enhanced
our appreciation of the diversity and polysemy of possession phenomena, the
turn away from explanatory endeavours altogether was overly hasty.
6
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
Explanatory theories should be generated from and consistent with the
relevant evidence available. In explanations of cultural phenomena, such
evidence nnneeeccceeessssssaaarrriiilllyyy includes data concerning the ‘structure, organization,
reproduction, and meaning’ of the phenomena. Insofar as these aspects of
possession demonstrate historically and cross-culturally recurrent features,
albeit manifested in culturally specifi c settings and ways, I suggest that the
identifi cation of general mechanisms at work is a worthy anthropological
endeavour that can mutually engage with descriptive and interpretive ap-
proaches. Specifi cally, in generating hypotheses about the incidence and
spread of recurrent aaannnddd variable spirit possession beliefs and practices, we must
consider both the ethnographic and scientifi c data on how such concepts and
beliefs are represented within and communicated among human minds.
Previous medicalist theories of possession and trance phenomena were
not only largely culture-blind; they were also mind-blind. Theories of how
possession concepts arise and persist are developed through observations
of how people conceptualize possession on the ground. These observations
and theories can guide how we carve up the whole domain, identifying like
forms of possession, not according to categories imposed from outside,
but according to categories produced by natural cognition. It is necessarily
through attention to the details and complexities of how people represent
possession on the ground that we can develop explanatory theory and hence
working categories and defi nitions for the comparison of truly similar cultural
phenomena. The following account of mental processes entailed in the ge-
neration and persistence of possession concepts demonstrates the need for a
categorical distinction within the broad range of phenomena that are either
lumped together in the scholarly literature as ‘possession’ or categorized on
arbitrary and superfi cial grounds. Insofar as the cognitive mechanisms identifi ed
represent panhuman capacities, the categories developed should not require
any ‘imposition’ on the cultural phenomena we seek to understand.
Possession Concepts
Below I offer some tentative suggestions in an effort to begin to develop
at least partial, provisional answers to the question of how possession con-
cepts are cognitively represented. I consider the profi les of two widespread
conceptual forms. These forms are preliminarily distinguished and identifi ed
according to the basic causal structures that characterize the various posses-
sion concepts. The fi rst form appears to entail recognizably recurrent pat-
terns of thinking to do with spirits as entities that cause illness. The second
What is Spirit Possession? 7
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
cross-culturally recurrent set of concepts entails causal reasoning to do with
the relationships between persons and bodies. After a brief introduction to
the these concepts, I consider how these different conceptual structures are
underpinned by (at least partially) different sets of cognitive mechanisms
and processes and how the relevant cognitive mechanisms serve to constrain
their variation and transmission.
In her account of spirit possession in Ghana, Margaret J. Field gives a clear
description of certain features of the local possession concept. Her remarks
concern the Ga ceremony of ‘driving away a bad gheshi ’. Field writes, ‘A bad
ggghhheeessshhhiii is thought of as an infl uence of ill-luck inhabiting a person and bringing
him misfortune and unprosperity when circumstances appear propitious. For
the expulsion procedure he is taken by the medicine-man and his apprentices
to the outskirts of the town and the bad ggghhheeessshhhiii driven out of him into the
bush or perhaps tied to a post or even induced to enter a fowl which is then
driven away’ (1969 : 11–12). In Susan M. Kenyon’s account of zzzaaarrr spirit pos-
session activities in Central Sudan, she describes the case of a woman called
Amna with a condition the woman described as ‘a beating in my stomach,
headaches, pains, cramps and vomiting . . . vomiting until I fainted’ (1999 : 94).
Kenyon relates how, after various diagnoses, treatments and setbacks over
a period of years, Amna reluctantly entertained the possibility that she was
possessed by zar spirits. Only when she was close to death did she perform
the costly kkkaaarrraaammmaaa ceremony in order to meet the demands of the possessing
spirits. Following the fi nal ritual procedure, Amna recovered from her long
illness. She confi dently attributed her recovery to her performance of the
karama ceremony. Janet McIntosh describes Giriama possession, in which
spirits ‘may spontaneously possess an individual, making him or her fall ill
or ruining their fortunes’ (2004 : 100). On confi rming via divination that pos-
session is the cause of illness or misfortune, the spirits are either exorcized
or mollifi ed through the host’s conversion to Islam. Within the context of
‘peripheral possession’, in which predominantly women are ‘affl icted’ with
possession, Ioan M. Lewis states that possession is ‘diagnosed and treated
as illness’ (1989 :79). The illness is often cured through exorcism of the in-
trusive spirit or accommodated through the performance of ceremonies that
‘tame’ the spirit. A catalogue of such possession-as-illness concepts is given,
including numerous examples from Africa, South America, and South-East
Asia (ibid.: 64 –77).
Although drawn from across many different cultural contexts, these va-
rious descriptions of possession manifest important similarities. Principally,
8
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
they commonly involve the attribution of misfortune, such as illness, to the
intrusion of a spirit into a person’s body — an extremely widespread notion
cross-culturally. Before we consider more fully the recurrent features of
such possession concepts, and the cognitive mechanisms underpinning their
structure, let us briefl y look at some further descriptions of possession, in
which the effects of the possessing spirit appear to be reasoned about quite
differently.
In many forms of possession concept, a spirit is represented as entering a
person’s body and as displacing or eclipsing the agency (or mind or soul or
spirit, etc.) of the host, thereby causing a change of identity. For the duration
of the episode, the spirit is said to be responsible for all behaviour and speech.
Melville Herskovits writes, ‘The individual thereupon is held to be the deity
himself’ (1948:66). In Mayotte possession, according to Lambek, ‘Spirits enter
the bodies of human beings and rise to their heads, taking temporary control
of all bodily and mental functions . . . The emphasis is on the ccchhhaaannngggeee that has
occurred. Despite the fact that the body remains the same, it is now occupied
by a different person (1981: 40). Paul Stoller writes of Songhay possession,
‘Spirit mediumship results from the temporary displacement of a person’s
double by the force of a particular spirit… The medium’s body has become
a deity’ (1989 : 31).1 The same fundamental concept is described by Dan
Rosengren in his account of human-spirit interaction in the Amazon; ‘The
general understanding, it seems, is that the soul of the shaman leaves while
the spirit enters his body. This metempsychosis is possible since the body is
something exchangeable which can be worn by various users; metaphorically
it is described as the dress, imanchake, of the soul’ (2006 : 812).
During my fi eldwork with a population of Afro-Brazilian religionists in the
northern Brazilian city of Belém, the majority of research participants offered
similar descriptions of possession. A senior member of the group described
possession as occurring when the body of the medium joins with the spirit
entity. These two parts, he claimed, make up the new (possessed) person.
Another senior member described possession as the moment in which one’s
own spirit withdraws ‘and another spirit comes and throws him/herself into
your body’. Drawing a clear demarcation between herself and the possessing
spirit, another member described her possession episodes as follows: ‘I don’t
know where my spirit goes. I don’t know. I only know that I switch off. I
don’t remain in me.’ When one is possessed, one’s own spirit is said to ‘lie
down’, ‘journey to the other world’, ‘dream’, ‘sleep’, or ‘remain watching’.
The spirit entity is said to ‘take control’, ‘dominate the mind’, or ‘command
What is Spirit Possession? 9
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
the body and the mind’. As one person stated, ‘Possession for me is a state
of unconsciousness . . . in which we are not answerable for our actions, our
bodily movements . . . we don’t have control of our bodies anymore. It’s the
total loss of control of the body and the mind. Something else controls — it
is the spiritual being’.
Such possession concepts, which I propose to label ‘executive possession’
concepts, may be defi ned as minimally entailing the following features: (a)
the presence of an incorporeal intentional agent in or on a person’s body,
that (b) temporarily affects the ousting, eclipsing or mediation of the person’s
agency and control over behaviour, such that (c) the host’s actions are partly
or wholly attributable to the intentions, beliefs, desires and dispositions of the
possessing agent for the duration of the episode. These features collectively
represent the basic causal structure of executive possession.
In contrast, the fi rst form possession concept described above – ‘patho-
genic possession’ – minimally entails the following set of conditions: (a) the
presence of an agent in or on a person, that (b) either causes no (perceived)
effects (i.e. the spirit is ‘dormant’)2 or causes physical effects, such as disease
or illness, or psychological effects, such as depression or hallucinations, or
existential effects more broadly defi ned, such as fi nancial misfortune, and
that (c) may persist indefi nitely or until a diagnosis is made and the agent is
dispossessed of the host’s body. This concept does not entail the displace-
ment of the person’s identity. It does not require, for example, that the person
is addressed by a different name — the name of the possessing agent — as is
commonly the case in executive possession episodes. The spirit’s name is
often not known until steps are taken to eliminate or mollify it. Indeed, ritual
naming ceremonies frequently appear to concern the establishment of the
possessing agent’s identity as a person, no longer an unknown, unpredictable,
and unbiddable thing or force.
There have been various attempts in numerous infl uential sociological
and anthropological theories to capture the general principles common to
‘pathogenic’ possession. Lewis claimed, for example, that this form of pos-
session was defi nitive of ‘peripheral’ cult activity, the domain of oppressed
women of marginal social status, and an expression of protest and resistance.
Lewis’s theory, perhaps the last ‘master narrative’ of any potential explanatory
import, came in for wide criticism partly on the grounds that its predictions
about marginality, oppression and resistance failed to resonate with a wide
range of cases (see Cohen 2007a: 69). Although there has been little further
consideration of the possible factors shaping and constraining the range of
10
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
forms that spirit possession concepts take, and the cross-cultural transmission,
persistence and patterns of distribution of these forms, the same patterns are
still implicitly recognized. McIntosh, for example, draws attention to some
of the key features of such possession ideas and practices in contrast to other
forms. She writes, ‘Unlike the short-lived trance possession brought about
by drumming and dancing during diviners’ spirit-propitiation rituals, pos-
session by coercive Muslim spirits is a chronic state in which hosts go about
their lives in a state of ordinary awareness while their bodies and actions are
subject to the spirit’s intervention’ (2004: 101). Lurking in this description is
a lingering differentiation between possession and ‘trance possession’ forms
that turns on the presence or absence of trance (see also Budden 2003 : 31–32).
I suggest, however, that insofar as trance is apparently present in one form
and not in the other, this is yet one further feature of the phenomenon to
be explained.
It is important to note that the forms of possession preliminarily defi ned
here do not refer to types of hosts or mediums, or cults or religious systems,
or even possession episodes. For example, a person who is conceptualized as
experiencing pathogenic possession, as defi ned above, may be possessed by an
intrusive agent that causes a range of undesirable physical symptoms, such as
nausea and dizziness. The same person may undergo what is conceptualized
as executive possession as part of the prescribed procedure for ascertaining
the identity and demands of the possessing agent. If elimination of the agent
follows, the person may be conceptualized as being no longer possessed. This
preliminary analytical distinction, therefore, is not premised on classifi ca-
tory systems, as was frequently the case with medicalist and sociological
theories, that box types of participants or possession cults together in ways
that necessitate the hardening, fi xing and artifi cial naturalization of culturally
supple and dynamic phenomena. It parses distinct cultural phenomena at
the level of cognitive representation, a level at which the cognitive sciences
aaarrreee now identifying a degree of rigidity and constraint, a level at which
there are potentially ascertainable material causes and effects upon cultural
transmission. To consider how this occurs, let us now turn to the cognitive
processes that are activated by these conceptual forms and that give rise to
their distinct causal structures.
The Cognition of Possession: Executive Possession
Executive possession concepts employ a vast range of cognitive systems,
many of which work below the level of conscious awareness. The majority
What is Spirit Possession? 11
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
of the inferences delivered by these systems go undetected, enabling people
to negotiate the mechanics, commands, and simple problems of everyday
life effortlessly and automatically. For example, the identifi cation of persons
we know well, from one day to the next and from one year to the next, is, on
the whole, an effortless process. In fact, there are certain cognitive processes,
dedicated to solving specifi c kinds of problems, that function independently
of and even resist conscious control. Recognising faces is one such process.
Those of us whose face-identifi cation systems are intact could make every
effort to imagine we are prosopagnosic (i.e. face-blind) while watching the
news or fl icking through our fi eldwork photographs, or even to imagine
what it must be like to be prosopagnosic, but we would fail quite miserably.
We could not help but identify a face when we saw one, and identify the
specifi c persons familiar to us when observing their faces, regardless of how
much we consciously endeavoured to switch off such processes. Similarly
tenacious systems constrain how we tacitly construe the nature of personal
identity, and therefore the nature of identity transformation. These systems
are at the core of executive possession concepts.
Erika Bourguignon noted that ‘the concept of spirit possession is clearly
dependent . . . on the possibility of separating the self into one or more ele-
ments’ (1968 : 4). More precisely, executive possession concepts frequently
entail a (literal or effective) separation of person from body. For example,
the agency of the host is often represented as withdrawing from the body
or assuming a passive role in relation to the control of the body, which is
subsequently occupied or animated by the possessing agent. Recent research
in developmental psychology suggests that person-identity is underpinned
by a dualistic distinction between oneself and one’s body that emerges early
in childhood (Bloom 2004). 3 Psychologist Paul Bloom refers to this as com-
mon-sense, or intuitive, dualism. His argument, drawn from a growing body
of experimental research as well as developmental and evolutionary consid-
erations, has generalizable, cross-cultural implications (see Gopnik, Meltzoff
& Kuhl 1999; Sperber, Premack & Premack 1995). Dualism, he claims, emer-
ges as an evolutionary by-product of the fact that humans have two parallel
cognitive systems engaged in the perception of bodies and persons. These
systems, known as ‘folk-physics’ and ‘folk-psychology’, are the focus of an
extensive literature in the psychological sciences. One system deals with the
physical world through the application of intuitive theories about push, pull,
cohesion, contact, and so on. The other system deals with the social world
and applies causal theories that have to do with mental attributes, such as
12
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
beliefs, desires, dispositions and intentions. Bloom suggests that these systems
deliver incommensurable outputs. As a result, humans are represented both
as psychological agents and physical objects – but these two co-existing sets
of cognitive representations never fully achieve coherent integration. This
gives rise to a dualistic perception of ourselves and others.
There is compelling anecdotal evidence and growing experimental evidence
in support of this position (e.g. Kuhlmeier et al. 2004). The failure to integrate
body-relevant outputs and person-relevant outputs of these distinct systems
is what leads children to suggest, for example, that their brain helps them do
maths, but that ttthhheeeyyy love their sister or brother (Bloom 2004). Bloom writes,
‘Our intuitive dualism grounds our understanding of personal identity. We
recognize that a person’s body will age; it might grow or shrink, lose a limb,
undergo plastic surgery — but in an important sense, the person remains the
same’ (2004 : 195). Intuitive person-body dualism underpins notions that one
can survive one’s biological death. It supports representations of disembodied
spirits as well as, for example, fanciful musings about combining ttthhhaaattt person’s
mind in that person’s body (for example, to create the perfect partner).
Of course, there are many situations in which the fact of the ‘embodiment’
of person-identity is highly signifi cant. Humans identify other persons by
paying special attention to the face or voice; we readily appraise persons
according to their age, sex, and skin colour (Hirschfeld 1999); we posit in-
timate associations between who we are — our self-essence — and particular
body parts, such as the brain, heart, stomach or blood, and so on. There is
some indication, then, that in the course of development, children and adults
acquire a deeper understanding of the interdependencies of body and person
or self and apply this in reasoning about the social world at least in some
contexts some of the time. Nevertheless, there are many situations in which
the interdependencies of self and body appear to be denied. Adults com-
monly distinguish between the things that they consider ttthhheeeyyy do and what
their bbbooodddyyy does, for example in concepts of death and afterlife, and even in
the more everyday recognition, for example, that their body lets them down,
or belies their ‘mental age’, and so on.
Many people explicitly deny holding a strong Cartesian view of the world.
On the materialist view, for example, all thoughts, memories, passion, love,
as well as mathematical and artistic competence are expressions of brain ac-
tivity. It is suggested, however, that even materialists can be iiinnntttuuuiiitttiiivvveee dualists
— their explicit and often elaborate and complex theories largely fail to pene-
trate or inhibit the activation of their ancient, automatic, and rapid cognitive
What is Spirit Possession? 13
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
processors. Materialism presents diffi culties that parallel those experienced
when one attempts to simulate prosopagnosia. Even those who know nothing
of brains, who have no explicit or lexicalized representation of self- or per-
son-identity, can be iiinnntttuuuiiitttiiivvveee dualists, such as when they posit the existence
of the self-essence (or, for example, soul) after death or the displacement of
the self from the body in a possession episode. There is considerable scope
for the systematic investigation of cross-cultural variation and recurrence
in the precise conditions that activate ‘embodied self’ intuitions and the
conditions that activate ‘self-body dualist’ intuitions in thinking about the
social world.
Although this fi eld of research is still young, the theories and fi ndings
developed thus far, based on experimental, observational and cross-cultural
ethnographic research support the conjecture that the cognitive capacity to
grasp and employ concepts that represent the autonomy of person-identity and
body is widespread, if not universal, and emerges early in childhood (Bering &
Bjorklund 2004; Bering, Hernández-Blasi & Bjorklund 2005; Bering & Parker
2006). The cognitive mechanisms that underpin the panhuman capacities to
represent persons as having an identity, and that are mobilized in identifying
other persons in our environment as in some important way the same person
from one day to the next, and from one year to the next, and that guide our
intuitions about the relations between person and body also appear to be
mobilized by executive possession concepts. These mechanisms support,
constrain and organize the ways in which possession is represented.
In keeping with intuitive expectations about person-identity, for example,
Spirit x is Spirit x whether he possesses Host a today or tomorrow, and
whether he is possessing Host a or Host b. When possessed, the host’s body
is represented as containing a different person-identity, that of the possessing
spirit. The new spirit-host entity may have the aaappppppeeeaaarrraaannnccceee of the host, but is
identifi ed and addressed as the spirit. Proper names are replaced, old con-
versations are taken up from where they left off previously, and interlocutor
and observer endeavour to apply the appropriate identity fi le to the entity
that now stands before them. The semantic and affective information, im-
pressions, and assessments of the possessing agent, contained in the ‘identity
fi le’, together with basic folk-psychological capacities, enable the observer
coherently to interpret the entity’s behaviours and utterances. Alterations
of voice and vocabulary, special abilities to heal and counsel, and apparently
miraculous feats are now explainable against this background of information
about the person-identity of the possessing agent.
14
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
Executive possession concepts typically entail the effective or literal
separation of person from body and the temporary establishment of a new
person-body confi guration. Such concepts subscribe to a radical form of
person-body dualism. As such, they are supported by ancient, effi cient, and
powerful mechanisms that underpin the capacities to represent our social
and physical worlds according to different sets of principles, and to represent
people as having identities that are fundamentally continuous, and that are
readily conceptualized under certain circumstances as quite separate from
the body and persistent in the face of bodily transformations and even death.
It is in terms of these cognitive tools that the common principles defi ning
executive possession concepts may be constructively characterized. It is also
in terms of these tools that we can appreciate the crucial distinction between
executive and pathogenic possession.
The Cognition of Possession: Pathogenic Possession
The distinction betwen the two forms may be crudely understood as fol-
lows: pathogenic possession concepts primarily concern the incorporation
of spirit-as-essence, not spirit-as-person, into the body. Although the identity
of the particular spirit, or class of spirits, is commonly identifi ed at some
point in the diagnosis or curing/socialization/exorcism process, the spirit
is primarily and most basically represented as a contaminating substance or
essence (material or immaterial). It is often agentized only secondarily, i.e.
it is represented as having thoughts and desires and acting in accordance
with goals. This representation, however, does not entail the displacement
or transformation of person-identity, as in executive possession. Pathogenic
possession, therefore, is conceptualized very differently from executive pos-
session, and employs a (partially) different battery of cognitive tools. Below
I suggest that this form of possession is guided by cognitive mechanisms that
deal with the representation of contamination and illness.
Beliefs about contamination are universal. The range of content expressed
in these beliefs and the fears and responses elicited by them exhibit wide
inter-individual and cross-cultural variability. The fundamental principles
according to which these beliefs are applied, however, demonstrate consid-
erable regularity across all contamination contexts (e.g. principles to do with
avoidance, contact, and purifi cation). Many of these principles are in place
by early childhood, forming part of young children’s understanding of illness.
Much of the research on this area in developmental psychology is concerned
with the question of whether young children have a biological representation
What is Spirit Possession? 15
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
of the transmission of contagious diseases. Research over the last two decades
indicates that young children are sensitive to contamination and that they
hold a ‘skeletal framework level’ understanding of contamination (Raman
& Gelman 2005 : 172). A fundamental principle organizing their theories of
contamination concerns contact between contaminants and uncontaminated
substances. Young children also understand that contamination may occur
through invisible mechanisms (e.g. through the transmission of germs). Further
fi ndings suggest, however, that young children have little knowledge of the
biological nature of such entities and the biological mechanisms by which they
multiply and cause illness. They fail to grasp the underlying mechanisms of
transmission and do not differentiate the contaminating processes of poisons,
germs and irritants (see Hejmadi, Rozin & Siegal 2004).
Nevertheless, by 4 years of age children have a functional understanding
of what sorts of things in the world are regarded by their elders as potential
contamination threats and a hyper-effi cient emotional (and therefore behav-
ioural) response to such potentially harmful entities. Contaminants evoke
strong fear and disgust, or revulsion, responses. Evolutionary psychologists
have suggested that such disgust emotions evolved in response to evolutionary
pressures to avoid survival-threatening contaminants (e.g. Haidt et al. 1997).
These responses are so powerful that they often persist despite the removal
of all traces of contamination — as Hejmadi, Rozin and Siegal found in their
cross-cultural study of Hindu Indian and American children’s responses to
contamination and conceptions of purifi cation, ‘once in contact, always in
contact’ (2004). The fi rm principles concerning contact and contamination,
and associated emotional responses, emerge early in childhood and form the
basis of full adult understandings about contaminants and how to avoid and
eliminate them. I suggest that they are also central to successfully spreading
pathogenic possession concepts.
Cross-culturally recurrent possession concepts involving the incorpora-
tion of agents into the body frequently occupy a place in local aetiologies
of disease. Notions about spirit intrusion and infestation and possession
epidemics parallel notions about the incorporation of poisonous substances,
the ingestion of rotten foods, the contraction of contagious illness, or the
inheritance of witchcraft substance. The basic concept resembles concepts
of disease and illness the world over, in that it involves a causal structure
that links cause (immediate and secondary), symptoms, prevention (e.g. by
means of avoidance) and cure (e.g. by means of expulsion and/or cleansing).
The details of the specifi c mechanisms by which spirits enter the body and
16
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
interact with the host’s biological processes are rarely articulated (or, at least,
are rarely reported in the anthropological literature). Empirical observations
of the kinds of symptoms that may be diagnostic, the causes to which they
may be attributed, and the (ritual) procedures by which they may be treated
are readily integrated into a compelling theory without requirement for
recourse to underlying mechanisms.
For example, the incorporation of the possessing agent may occur due to
weak spiritual or physical defences — defences should therefore be fortifi ed.
‘Vaccination’ procedures, aimed at barring the entry of these spirits, may be
encountered throughout the literature. For example, in certain Afro-Brazilian
rrreeellliiigggiiiooouuusss gggrrrooouuupppsss,,, ‘‘‘ttthhheee llloooccckkkiiinnnggg ooofff ttthhheee bbbooodddyyy’’’ rrriiitttuuuaaalll (((fffeeeccchhhaaammmeeennntttooo dddeee cccooorrrpppooo) is be-
lieved to guard it from the entry of bad energies. Further, possessing agents
are often thought to be ingested in food, or they may pass into the body
through any of its orifi ces. Possession is often believed to be the result of
association with other possessed persons. Such persons should therefore be
avoided. In the case of Amna (mentioned earlier), we learn that she believed
that she had not inherited the zzzaaarrr from her mother, but rather she had caught
it from a possessed neighbour whom she had visited. The very language used
here is consistent with the possession-as-contamination conceptualization.
The spirit gets under the skin, so to speak, as do germs or poison, and causes
physical and psychological maladies. The minimal concept of possession is
then often afforded a much richer, interconnected structure with its wider
social and functional signifi cance. Why Amna, and why at that specifi c point
in her life, for example, are deeper, existential questions readily answerable
in terms of the principles of social exchange. The notion of an intentional
agent with special powers to exert such effects over Amna’s body becomes
particularly salient. The spirit’s demands must be met, and social debts clear-
ed in exchange for the restoration of health and control. In other possession
contexts, the spirit may be eliminated from the body, thereby restoring the
body to a decontaminated and healthy state. Elaborate ritual procedures are
employed to vaccinate people against such intrusions, or to ‘bind’ the activities
of the possessing spirits, or to expel and purge them from the body.
The Transmission of Possession Concepts
The support that certain possession concepts receive from natural and
early-emerging cognitive capacities enhances their memorability, communi-
cability, relevance and, therefore, their potential incidence in the cross-cultural
record. It appears that people the world over have little diffi culty grasping
What is Spirit Possession? 17
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
and applying concepts of executive possession and related concepts that rest
on similar underlying assumptions (e.g. to do with the continuity of person-
identity after biological death). Indeed, if we break free from the traditional,
common sense defi nition of possession as something necessarily involving
other-worldly spirits, and frequently occurring in what may be considered
‘religious’ contexts, we encounter executive possession in a wide variety of
contexts, including novels and fi lms (e.g. DDDrrr MMMaaabbbuuussseee aaannnddd SSScccoootttlllaaannnddd YYYaaarrrddd). I sug-
gest that these concepts spread successfully because they are supported by
panhuman mental capacities that are employed in the resolution of everyday,
common problems. Some of these capacities, such as distinguishing between
non-agentive objects, such as a bedpost, and intentional, psychological entities,
such as mother, are being exercised quite literally from birth (see Gopnik,
Meltzoff & Kuhl 1999).
It is important to note, however, that the very same cognitive mecha-
nisms that facilitate the emergence and spread of cross-culturally wide-
spread executive possession concepts also serve to constrain the potential
variability of such conceptual forms in cultural transmission. It appears, for
example, that although there are numerous logically possible confi gurations
and combinations of spirits and hosts that follow the conceptual structure
of executive possession, there is one in particular that appears to have a
higher incidence in the cross-cultural record than any other. This concept
entails the complete displacement of the host’s agency by that of the spirit.
Alternative executive possession concepts (e.g. entailing the merging of host
agency and spirit agency in the host’s body) do arise in the ethnographic
record, but are rare and appear to require extra cultural scaffolding in order
to facilitate their transmission (e.g. repetitive instruction, rehearsal) (Cohen
2007a). Preliminary experimental and ethnographic analysis suggests that
‘displacement‘ concepts better exploit natural cognitive dispositions and
tendencies than ‘fusion’ concepts and other variants (see Cohen & Barrett,
in press a; Cohen & Barrett, in press b). Further systematic, cross-cultural
research is required, however, in order to identify more precisely how the
contours of person-body reasoning inform the variable emergence and spread
of different forms of executive possession concepts.
Likewise, because pathogenic possession notions are supported by early
emerging predispositions to acquire certain kinds of concepts concerning
contact, contamination, prevention and purifi cation, I suggest that those forms
of pathogenic possession that do not capitalize upon these intuitions will be
more diffi cult to grasp, recall and communicate than those that do. Consider,
18
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
for example, the transmission potential of the notion that the weaker one’s
spiritual immune system, the more prepared it is to resist the possessing
agent’s entry into the body. Or that preventative measures should entail the
ingestion of spirit-inhabited substances. Such concepts run counter to the
intuitions spontaneously delivered by our cognitive systems. Without recourse
to enriched learning conditions, concepts that deviate from the basic structure
of contamination concepts are unlikely to enjoy widespread success.
Further, pathogenic possession concepts that specify elaborate theories of
precisely how the contaminant works its effects may also require considera-
ble cultural support. A rare example of such an elaborate theory is described
by Nancy Caciola in her discussion of spirit possession in medieval Europe
(2000). Caciola describes how the medieval Catholic Church developed
a model by which divine inspiration (or possession by God or the Christ)
could be distinguished from demonic possession. The model was labelled the
‘discernment of spirits’ and was the means by which ‘medieval intellectuals
attempted to naturalize the discernment process by elaborating a physiolo-
gical theory that differentiated the precise, internal mechanisms of divine
from demonic possession’ (2000 : 272).
The discernment model made extensive use of medieval medical under-
standings of the human spirit, or spiritus. It was believed that the main seat
of the spirit was the heart, but it was thought to pervade the body and to
compose three different categories — the vital spirit, the natural spirit, and
the animal spirit. Demons, it was proposed, enter the body, interfering with
the functions regulated by the natural and animal spirits, such as digestion,
heartbeat and respiration. Only the Holy Spirit entered the heart, the seat of
the soul, replacing or joining with the human spirit. This latter concept ap-
pears to resonate more strongly with the features characteristic of executive
possession. Indeed, this is arguably what motivated the elaboration of such
theories in the fi rst place — the need to know whether one is interacting
effectively with God (a new person-identity) or with a deranged individual
whose demons do not replace the soul but rather ‘tempt or confuse’ it.
This example demonstrates how special socio-cultural conditions facili-
tated the emergence and transmission of exact theories about the internal
mechanisms of possession contamination — in this case, such conditions
included the knowledge and integration of theological and medical theory,
the presence of an ecclesiastical orthodoxy and elite, the faithful transmission
of the discernment model by means of texts and sermons, and the rigorous,
principled adjudication of the details of each case according to the conven-
What is Spirit Possession? 19
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
tions and formulas of the model. Factors such as these considerably enhance
the transmissive potential of concepts that elaborate considerably upon the
basic intuitions delivered by our cognitive systems.
This brings us to some fi nal reiterations and clarifi cations of the distinc-
tion between pathogenic and executive possession, and of the purpose of
developing the present approach. First, executive possession does not turn
on considerations of whether trance is present or absent. Trance-free divine
inspiration, brought about by the incorporation of God into one’s body, may
be cognitively represented in much the same way as executive possession
involving dissociated states. It entails the temporary replacement of agency
and the attribution of utterances to the possessing entity. This minimal
defi nition does not assume or necessitate an alteration in the conscious
state of the host. The host could be partially or fully conscious of the situa-
tion but unable to exert any executive control over his or her behaviour. It
is not trance, therefore, that is a central feature of the conceptual structure
of this representation of possession, but the change of agency and identity.
Possession concepts among the Bigajos Islanders entail transformations of
core person-identity without trance — ‘defunct women’ are possessed by the
spirits of uninitiated warriors and undergo posthumous initiation aaasss these
men, enabling the men to reach their ancestors’ land. The defuncts ‘manifest
the fi ery warriors that possess them’ — ‘wives and mothers are now warriors’
(de Sousa 1999 : 85). Some women show a trance-like behaviour, but this is
not a necessary component of the perceived identity transformation. Trance,
therefore, is a feature that is frequently associated with, but not causally ne-
cessitated by, the perceived temporary resignation of the host’s agency and
executive control. Furthermore, the concept of pathogenic possession does
not preclude the possibility of trance states. In fact, episodes of dissociation
and disorientation are regularly recurring psychological effects of this form
of possession.
Second, pathogenic possession need not only concern negative contami-
nation. Although much rarer in the ethnographic literature, some possession
concepts entail the incorporation of agents that are perceived to cause desirable
effects (e.g. fruits of the Spirit). Some agents may be represented as possessing
the body but as having no, or neutral, effects. I suggest that theories by which
negative, positive, and neutral agents enter the body are guided by a similar
set of principles concerning contact and elimination, in much the same way
as the ingestion of poisons, antidotes, nutrients and placebos are considered
in terms of assimilation and effects. They are, however, accompanied by
20
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
very different sets of emotions. The highly salient and persistent negative
emotions associated with negative contamination and the associated threat
to personal well-being, survival and control may be key factors contributing
to the relatively higher incidence of negative pathogenic possession concepts
cross-culturally. Such threatening, negative, personally consequential situa-
tions are more attention-demanding than those involving positive effects.
Ultimately, where attention is not given — for example, when ritual curing is
not performed — the host may even face death. Such potential consequences
create heightened alertness to the possibility of pathogenic possession, where
such interpretations are already considered plausible, and increased demand
for vaccination, purifi cation and elimination procedures.
Third, whether possession may be identifi ed as pathogenic or executive
necessarily takes into consideration only the fundamental features of the
conceptual structure of the possession representation. The distinction is not
one between consciousness and dissociation, male host or female host, path-
ological host or healthy host, good spirit or evil spirit, or ecstatic or sedate
state. A person may be pathogenically possessed by a particular spirit at one
moment, and executively possessed by the spirit at another, depending on
whether the perceptual and conceptual inputs match the conditions that
activate contamination systems or person-identity systems in cognition.
Parsing the domain at the level of conceptual structure leads to an inter-
esting fi nal observation. There is considerable common ground between
pathogenic and executive possession. As stated previously, both concepts
typically entail a representation of incorporeal agents entering the host’s body
and working various kinds of physical and mental effects. These phenomena
may demonstrate considerably greater continuity, however, in terms of their
basic cognitive representation and conceptual structure, with other cultural
phenomena that are not typically considered within the common-sense cate-
gory of possession. Pathogenic possession concepts, for example, appear to
display considerable overlap with certain forms of concepts about witches and
witchcraft essences (Cohen 2007b). It has been widely noted by ethnographers
that the potential to perform witchcraft is often believed to have its source
in a substance or ethereal essence housed within the body of the witch. In
the famous case of the Azande, the substance may be inherited from father
to son and from mother to daughter (Evans-Pritchard 1976); in the case of
Akan it may be caught through contact with objects onto which witches
have transferred their witchcraft (Debrunner 1961). Only the removal of the
essence or substance guarantees the elimination of the witchcraft potential.
What is Spirit Possession? 21
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
In these examples, the parallels with pathogenic possession and with notions
about contamination are clearly evident. Executive possession concepts,
in turn, display signifi cant continuities with concepts of reincarnation (as
well as lay-understandings of Dissociative Identity Disorder). Both cultural
forms share the similar basic cognitive underpinnings and can be grasped
and understood according to similar sets of intuitive principles concerning
persons and bodies and the relations between them.
Conclusion
Why parse the domain of possession, and the cultural domain more broadly,
in terms of underlying cognition? In order to explain patterns of incidence of
cultural phenomena, we need to defi ne clearly the parameters and content of
the phenomena to be explained. Previous explanatory approaches to possession
employed common-sense-derived intuition to do this, producing polythetic
and ill-fi tting categories, based on arbitrarily drawn, culturally specifi c mea-
sures and criteria. Medicalist perspectives, too, were unsatisfactory because
they failed to take into account ‘structure, organization, reproduction, and
meaning’ of the phenomena under question. These aspects are critical to the
development of theories about why cross-culturally recurrent possession
concepts take the form they do and about what permits their transmission
and persistence across diverse cultural contexts and throughout history.
Approaches to these questions have generally started by identifying what
appear to be recurrent features, and arbitrarily selecting possible underlying
principles that might give rise to these distinctive patterns. To avoid the arbi-
trariness and randomness of such approaches, and to develop a plausible theory
of why possession concepts take the forms they do, I argue that it makes sense
to integrate fi ne-grained descriptive analysis with what we now know about
how human thought works. Cognition is a powerfully constraining factor on
the kinds of cultural concepts likely to emerge and persist. By identifying the
objects of our explanations in terms of their underlying cognitive mechanisms
and processes, we classify cultural phenomena, not, for example, according
to convictions delivered by common sense, or to aesthetically determined
preferences, or empirically impoverished fl ights of philosophical fancy, but
according to causally signifi cant criteria. We can then identify what kinds of
phenomena are alike and what are not in terms of these criteria. In the case
of possession, we can be clear about which forms and aspects of possession
our theories address, and what kinds of evidence are relevant (and which
are not) to investigating the theories. Ultimately, we can take steps toward
22
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
characterizing the interdependence of cognitive, ecological and contextual
factors shaping the transmission and cross-cultural incidence of recurrent
and variable forms of possession concepts and practices.
I have suggested that executive and pathogenic possession are differently
represented according to basic sets of panhuman cognitive processes. Exe-
cutive concepts of possession have at their core a natural conceptualization
of the world as composed of two discrete kinds of phenomena — physical
bodies that operate according to laws of physical forces (e.g. gravity, contact,
cohesion, etc.) and psychological agents that operate according to beliefs,
desires, dispositions, and so on. Recent research is beginning to show that the
perception of one’s self, or person, as distinct from one’s physical matter, or
body, may be less a product of a particular philosophical tradition, and more
the outcome of an intuitive dualist stance on the social world that originates
early in infant development (e.g. Bloom 2004). Readily transmittable execu-
tive possession concepts (such as ‘displacement’ concepts) capitalize upon
this natural, intuitive position. Widespread pathogenic concepts, in contrast,
capitalize upon cognitive systems that deal with problems of contamination
and, in many cases, illness. These different sets of cognitive systems and
processes are differently activated, make sense of different phenomena in
the social, physical and biological world, and mobilize different assumptions
and inferences about those phenomena. It is with increasing confi dence that
we can propose that (and how) the ways in which human cognition parses
the world have identifi able and material effects upon patterns of cultural
transmission.
Ultimately, why go back to the old debates about types of possession? I
suggest that anthropologists retired from these debates prematurely. Pos-
session, it was argued, is an explanation-defying, holistic social reality. ‘Its
province is meaning’, Boddy asserted, ‘and it is best addressed in that light’ in
terms of the ‘potential range of its signifi cance within the cultural and social
context’ (1989 : 136). Nevertheless, beneath the behaviours we observe (both
recurrent and variable) lie universal mechanisms and processes of cognition,
the workings of which our participants may be largely unaware. The com-
mon structures of this cognitive architecture impose appreciable constraints
on the forms that possession concepts (and other cultural phenomena) take.
Because there are features of unconscious cognition that ‘modify the proba-
bility’ (Bloch & Sperber 2002 : 729) that a concept will ‘catch on’ in cultural
transmission, there are therefore signifi cant questions to be formulated about
the role of basic cognition in the spread of possession phenomena. Ultimately,
What is Spirit Possession? 23
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
however, we need to balance explanatory and interpretive endeavours for
the development of theories that are sensitive to the broad range of factors
that shape the recurrent and variable structure, organization, reproduction,
and meaning of cultural phenomena. There is now considerable scope for
mutual engagement between cognitive science and anthropology. The timing
is perfect to revisit these issues in the scholarship on possession and to en-
tertain once again the possibility of articulating increasingly comprehensive
explanatory accounts of the complex and variable patterns of thinking and
behaviour so eloquently portrayed in the anthropological literature.
Acknowledgments
Thanks to Harvey Whitehouse and Justin Barrett for offering helpful suggestions on
an earlier version of this paper. It is in part the outcome of some very stimulating
discussions with Robert McCauley, to whom I also express my thanks and appre-
ciation.
Notes
1. The ‘double’ is described as ‘the essence of a person’s humanity’ and is distinguished
from the human body (which ‘consists of fl esh’) and the ‘life force’ (the ‘energy of
life’) (1989 : 31).
2. See, for example, Giaconda Belli’s novel, The Inhabited Woman.
3. Labels for what I am referring to as person-identity vary widely cross-culturally. In
some places, no such label may even exist. This variability does not constitute evi-
dence of conceptual variability, however, particularly at the level of tacit represen-
tation. Note that unless otherwise stated, the term ‘person’ and person-identity is
used interchangeably, but always in the strict sense described above.
References
Barrett, Justin L. 2004. WWhy Would Anyone Believe in God? Walnut Creek, CA: AltaMira
Why Would Anyone Believe in God?
hy Would Anyone Believe in God?
Press.
Belli, Giaconda. 1994. The Inhabited Woman. Wisconsin: University of Wisconsin Press.
Bering, Jesse M. & David F. Bjorklund. 2004. The Natural Emergence of Reasoning
About the Afterlife as a Developmental Regularity. Developmental Psychology, 40(2),
217 –233.
Bering, Jesse M., Carlos Hernández-Blasi & David F. Bjorklund. 2005. The Develop-
ment of ‘Afterlife’ Beliefs in Religiously and Secularly Schooled Children. British
Journal of Developmental Psychology, 23 : 587 – 607.
Bering, Jesse M. & Becky D. Parker. 2006. Children’s Attributions of Intentions to an
Invisible Agent. DDevelopmental Psychology, 42(2) : 253 – 262.
Developmental Psychology
evelopmental Psychology
Bloch, Maurice & Dan Sperber. 2002. Kinship and Evolved Psychological Dispositions:
The Mother’s Brother Controversy Reconsidered. Current Anthropology, 43 : 723–748.
Current Anthropology
Current Anthropology
Bloom, Paul. 2004. Descartes’ Baby: How the Science of Child Development Explains What
Makes Us Human. New York: Basic Books.
Boddy, Janice P. 1989. Wombs and Alien Spirits: Women, Men, and the Zar Cult in Northern
Sudan. New Directions in Anthropological Writing. Madison: University of Wis-
consin Press.
24
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
emma cohen
Bourguignon, Erika. 1968. A Cross-Cultural Study of Dissociational States. Columbus: Re-
search Foundation, Ohio State University.
—. 1976. Possession. Chandler & Sharp Series in Cross-cultural Themes. San Francisco:
Chandler & Sharp Publishers.
Boyer, Pascal. 2001. Religion Explained: The Evolutionary Origins of Religious Thought.
New York: Basic Books.
Budden, Ashwin. 2003. Pathologizing Possession: An Essay on Mind, Self, and Expe-
rience in Dissociation. Anthropology of Consciousness, 14(2) : 27 –59.
Anthropology of Consciousness
Anthropology of Consciousness
Caciola, Nancy. 2000. Mystics, Demoniacs, and the Physiology of Spirit Possession in
Medieval Europe. Comparative Studies in Society and History, 42 : 268 –306.
Comparative Studies in Society and History
Comparative Studies in Society and History
Cohen, Emma. 2007a. The Mind Possessed: The Cognition of Spirit Possession in an Afro-
Brazilian Religious Tradition. New York: Oxford University Press.
—. 2007b. Witchcraft and Sorcery. In Religion, Anthropology and Cognitive Sci-
ence, edited by H. Whitehouse & J. Laidlaw. Durham, NC: Carolina Academic Press.
ence
ence
Cohen, Emma & Justin Barrett. (in press a) When Minds Migrate: Conceptualizing
Spirit Possession. Journal of Cognition and Culture.
Journal of Cognition and Culture
Journal of Cognition and Culture
—. (in press b) Conceptualizing Spirit Possession: Ethnographic and Experimental Evi-
dence. Ethos.
Ethos
Ethos
Debrunner, Hans W. 1961. Witchcraft in Ghana: A Study on the Belief in Destructive Wit-
ches and its Effect on the Akan Tribes, Accra. Presbyterian Book Depot.
De Sousa, Alexandra O. 1999. Defunct Women. In Spirit Possession, Modernity and Power
in Africa, edited by H. Behrend & U. Luig. Oxford: James Curry Ltd.
in Africa
in Africa
Evans-Pritchard, Edward E. 1976 [1937]. Witchcraft, Oracles, and Magic Among the Azan-
de, abridged with an introduction by Eva Gillies. Oxford: Clarendon Press.
de
de
Field, Margaret J. 1968. Spirit Possession in Ghana. In Spirit Mediumship and Society in
Africa, edited by J. Beattie & J. Middleton. London: Routledge & Kegan Paul Ltd.
Africa
Africa
Gopnik, Alison, Andrew N. Meltzoff, & Patricia K. Kuhl. 1999. The Scientist in the Crib:
Minds, Brains, and How Children Learn. New York: William Morrow & Co.
Haidt, Jonathan, Paul Rozin, Clark McCauley, & Sumio Imada. 1997. Body, Psyche, and
Culture: The Relationship between Disgust and Morality. Psychology & Developing
Societies, 9 : 107 –131.
Hejmadi, Ahalya, Paul Rozin & Michael Siegal. 2004. Once in Contact, Always in Con-
tact: Contagious Essence and Conceptions of Purifi cation in American and Hindu
Indian Children. DDevelopmental Psychology, 40 : 467 –476.
Developmental Psychology
evelopmental Psychology
Herskovits, Melville. J. 1948. Man and His Works: The Science of Cultural Anthropology.
New York: A.A. Knopf.
Kenyon, Susan M. 1999. The Case of the Butcher’s Wife: Illness, Possession and Power
in Central Sudan. In Spirit Possession, Modernity and Power in Africa, edited by H.
Behrend & U. Luig, pp. 89 –108. Oxford: James Curry Ltd.
Klass, Morton. 2003. Mind Over Mind: The Anthropology and Psychology of Spirit Posses-
sion. Lanham, Md.: Rowman & Littlefi eld.
Kuhlmeier, Valerie, Paul Bloom & Karen Wynn. 2004. Do 5-month-old Infants see Hu-
mans as Material Objects? Cognition, 94 : 95 – 103.
Lambek, Michael. 1981. Human Spirits: A Cultural Account of Trance in Mayotte. Cam-
bridge Studies in Cultural Systems. Vol. 6. New York: Cambridge University Press.
—. 1989. From Disease to Discourse: Remarks on the Conceptualization of Trance and
Spirit Possession. In Altered States of Consciousness and Mental Health, edited by C. A.
What is Spirit Possession? 25
ethnos, vol. 73:1, march 2008 (pp. xx–xx)
Ward. Newbury Park: Sage Publications.
Lewis, Ioan M. 1971. Ecstatic Religion: An Anthropological Study of Spirit Possession and
Shamanism. Pelican Anthropology Library. Harmondsworth: Penguin Books.
Maurizio, Lisa. 1995. Anthropology and Spirit Possession: A Reconsideration of the
Pythia’s Role at Delphi. TThe Journal of Hellenic Studies, 115 : 69 – 86.
The Journal of Hellenic Studies
he Journal of Hellenic Studies
McIntosh, Janet. 2004. Reluctant Muslims: Embodied Hegemony and Moral Resistance
in a Giriama Spirit Possession Complex. JJJRRRAAAIII (NS), 10 : 91 – 112.
Raman, Lakshmi & Susan A. Gelman. 2005. Children’s Understanding of the Trans-
mission of Genetic Disorders and Contagious Diseases. DDevelopmental Psychology,
Developmental Psychology
evelopmental Psychology
41 : 171 – 182.
Rosengren, Dan. 2006. Transdimensional Relations: On Human-Spirit Interaction in
the Amazon. JJJRRRAAAIII (NS), 12 : 803 – 816.
Sperber, Dan, David Premack & Ann J. Premack. 1995. Causal Cognition: A Multidiscipli-
nary Debate. New York: Oxford University Press.
nary Debate
nary Debate
Stoller, Paul. 1989. Fusion of the Worlds: An Ethnography of Possession Among the Songhay
of Niger. Chicago: University of Chicago Press.
of Niger
of Niger