We usually think of order as confronting chaos, pitting structures and
plans against random unpredictability. But another approach is to distinguish
different types of order precisely in terms of how they interact with
indeterminacy. A type of order that was in some way open to indeterminacy
might learn to be more subtle and complex, with a wider range of possible
responses to the unexpected. A type of order that never interacted with
indeterminacy would, in contrast, stay fixed and closed. Whatever the
advantages of openness however, the open type of order clearly has a problem
which the closed type doesn't: how does the system ensure that the
input of indeterminacy doesn't directly erode, and even finally dissolve,
its own organisation?
Human cultures have, I suggest, adopted a specific solution to this problem:
openings to the indeterminate occur only at specific places and times,
or "phases", these being clearly distinguished from the other
more widespread phases during which indeterminacy is immediately assimilated
to determinate models. The cultural practices identified as religion and
art provide the main contexts within which these special phases happen.
However, religion and art also offer modes of retrospective integration
of the indeterminate, with religion typically re-presenting it as an expression
of universal order.
What they utilised of chance in divination practices was absolutely not
considered as such but as a mysterious web of signs, sent by the divinities....
(who were often contradictory but who knew what they wanted) and which
could be read by elect soothsayers.
Iannis Xenakis Towards a Philosophy of Music in Formalised Music.
From amongst the many types of phases for the interaction of order and
indeterminacy within human cultures, this article will single out shamanism.
I will argue that the technique of the shamanic trance is a method for
deliberately exposing the shaman to the aleatory within the human psyche
as a model or equivalent for the larger indeterminacies of the natural
Indeterminacy and Shamanism
My argument will partly build on, and partly depart from, what I consider
to be the single most important work of modern ethnography on Siberian
shamanism, Roberte Hamayon's La Chasse l'Ame. Towards the end
of this work, Hamayon sums her account of the functioning of shamanism
into the phrase la gestion de l'atoire. (roughly, the management
of indeterminacy, although gestion has less administrative connotations
in French than management in English). It is still possible within this
perspective to read the shaman's relations of exchange with the spirits,
expressed in alliance (for hunters), or filiation (for pastoralists),
as an interaction between order and indeterminacy. But for Hamayon the
act of shamanising, or conducting a shamanic seance, is no more nor less
than the symbolic exchange itself. Only in her conclusion does she retreat
from this unyieldingly semiological account of what happens in a shamanic
trance, to remind us that, if the sociology of shamanism can now be sketched
in - perhaps more than sketched for cultures where information is adequate - the
psychology of shamanism still waits to be written.
It is the hint of this opening left by Hamayon that I shall use to introduce
a distinct but complementary reading of the data - a reading based
on the notion that, although the indeterminacy which shamanism explicitly
addresses may be in the external environment, in the form of uncertain
food supplies (for hunters), or uncertain health (for pastoralists), the
act of shamanising activates the potential indeterminacy of the human
mind and is therefore not reducible to a symbolic exchange dependent on,
and conducted by, a continuously present and responsible narrative self.
Indeterminacy and Self
I draw from the work of Daniel Dennett the idea that the continuity of
the human conscious self is an illusion made necessary by a cultural need
for the continuous narrative projection and interaction of all members
of society. In fact, according to Dennett, in day to day life, consciousness
constantly suffers micro-lapses which it then papers over, so to speak,
to project to itself, and potentially to others, an appearance of ongoing
control. Much of the time, says Dennett, experience just happens: the
integral sense that is given to it is a retrospective construction, and
the all-powerful all-active decision-maker seated at the centre of the
human mind is simply an illusion.
The real matrix of experience is what he calls the parallel architecture
brain, or PAB. This is not an integrated structure with a central decision-making
core, but a cluster of many different kinds of modules, all with different
yet flexible modes of functioning, all having evolved in different evolutionary
epochs as responses to the changing demands of Darwinian evolution.
One implication of Dennett's account is that, if it is narrative
that defines the sense of self and is the essence of the human psyche's
auto-structuring process, then societies could, at least in theory, suggest
not only other narratives but other kinds of narrative. I propose that
types of symbolic exchange that putatively involve direct encounters with
other worlds, such as those of the spirits, will require a local and temporary
lapse in the normal social narrative. Dennett's unified narrative
self, or UNS, is not only an actor in this kind of exchange, but the bearer
of a symbolic value that is here given up and then returned. For this
to be the case, the relation between the UNS and the parallel architecture
brain has to be abnormal. The UNS has to enter a phase of temporary abeyance,
allowing in a lot more from the PAB, and only later reconstructing the
significance of the new material into a narrative. The micro-lapses of
everyday living that are usually constantly reabsorbed into the continuum
of the social narrative and its self, now become a continuous and prolonged
After the travel episode, the shaman sits down and starts telling stories
about what he has seen on his journey, and at the same time the spirits
repose who helped the shaman on his journey.
Triinn Ojamaa: The Shaman as the Zoomorphic Human
If this is so, then the problem of description shifts from the indeterminate
element in shamanic practice to the determinate element. That is: in what
way, and to what extent, is the information activated in the shamanic
trance organised? It is important to understand this question as not being
dissolved by the usual semiological or functionalist procedures which
assume a fundamental continuity between all types of practice occurring
in a culture. That is: the answer will be in terms of a type of articulation
between determinacy and indeterminacy strongly different from that proposed
for other types of social practice.
Even though a shamanic trance is an opening to indeterminacy, the trance
is evidently set in motion and brought to a conclusion by the use of determinate
ritual sequences that are carried over from one session to the next. This
is what makes shamanising a method, and distinguishes the trance from
an attack of madness. The elements of these sequences carry determinate
meanings. That is: the ritual is not the indeterminacy itself, but the
method for opening and closing a bounded zone of indeterminacy right inside
the ordered cultural system.
In fact I define ritual as a technique for separating out phases that
are normally intertwined and in mutual dynamic balance. On a psychological
level ritual reorganises the rhythm of experience, and, where it is used
for shamanic trance, this leads to episodes during which the narrative
of the self is postponed. Right now, however, I want to apply this definition
of ritual to the macro-level of the culture as an informational system.
On this level, ritual frames or separates out different phases in the
total informational process of the culture. It is not a completely different
category of actions but a set of formal changes in the informational aspect
of whatever objects, words or actions are brought into its sphere. These
formal changes have long been identified in anthropological literature
as exaggeration, stereotyping, and repetition. They represent a disruption
of the formal surface of functional communicative modes, and this corresponds
to an important shift in the relation between signifier and signified.
The terms "symbol" and "sign" can be used to define
the limits of the range of possible relations between signifier and signified.
The symbol is defined as having a highly present signifier, tending to
split into many parallel repetitive redundant intrances: singly or together,
these indicate a signified that remains fluid, absent, and relatively
undefined. Within the communicative field of the culture, the symbol then
allows certain limited operations involving a meaning that remains ambiguous
The sign, in contrast, has a signifier detached from any one manifestation
in time and space; its signifier is that aspect of the concrete thing
that can be abstracted, generalised, exchanged with another one similar.
Meanwhile the signified becomes less fluid, more fixed. So the sign spans
less of the possible distance between presence and absence than the symbol.
But it spans it more functionally because it enables the absent signified
to be configured as a distinct and determinate idea, and at the same time
one concrete situation to be compared with another.
It is not surprising then that there is a very concrete side to the way
in which ritual frames the shamanic trance. The ritual situates and frames
the trance in the now and the here. There is always the aspect of an attentive
re-situation and restitution of the participant(s) into the moment and
place in which they are present. The participants are deliberately withdrawing
from the mobility of the sign offered by the cultural system: they are
re-embedding the sign in a conspicuous tension between presence and absence.
And, as I suggest below, it may be a special concentration on the highly
present that triggers the shaman's imaginary absence or journey into
the domain of the spirits.
Symbols rather than signs also characterise the transmission of information
from older shamans to novices during training. This knowledge does not
explain how something works but involves familiarity with all the specific
occurrences of a phenomenon. Ritual activity requires repeating a procedure
until all the concrete intrances in a set have been exhausted. In some
forms of shamanic training every individual part of the initiate's
body must be individually "consecrated" by a spirit if the shaman
is not to risk death during healing seances.
The Training of a Siberian Shaman - Leonid Lar
And, as every ethnographer knows, interviews with shamans usually start
with the shaman displaying her or his knowledge in list form even if the
question was intended to avoid just that. Again what is important here
is not the "content" of the lists, the fact that each item can
be allocated a cultural meaning to be decoded or not decoded by the ethnographer,
but the repetitive parallelism of the form.
This distinct organisational character of ritual and symbolic information
shows that the nature of the determinate elements in shamanic practice
is such as to preclude their recuperation into the semiotic totality of
the culture. Such a recuperation is reductive because, judged as a sign,
a symbol is inefficient, ambiguous, and polyvalent, so that semiological
interpretation leaves out the main thing that symbols do, which is to
herald, activate, or refer back to, zones or phases of indeterminacy.
In some episodes of the shamanic trance we see the shaman acting out memorised
ritual sequences, whereas in others he or she appears physically disorganised,
or at least differently organised, and so incapable of intentional action,
and perhaps dependent on help from an assistant. If shamanising is nothing
else but symbolic exchange, this lack of control must be a theatrical
effect geared to a symbolic and communicative function. Are shamans then
just actors? Is the shamanic trance in fact a theatrical performance in
which the shaman pretends to communicate with spirits - presented as
autonomous and volatile - whilst actually enacting a symbolic exchange
according to the rules of that exchange so as to arrive at a predetermined
or otherwise determined result - the verdict, diagnosis, or healing?
There are parallels here with the current debate about hypnosis; do hypnotised
subjects just simulate being hypnotised or do they really enter a different
state of mind? The psychologist John Gruzelier identifies two main characteristics
of mental behaviour under hypnosis that indicate what can legitimately
be called a different state of mind. The first of these is that the brain
"turns in on itself", losing interest in sensations from the
external world and paying more attention to products of the imagination.
The second is that the brain stops testing, criticising, and verifying
perceptions; therefore products of the imagination become more credible.
I suggest that the shaman engages in partial self-hypnosis and that the
lapse in the UNS and opening towards the PAB is achieved via the inhibition
of both attention to the outer world and criticism and verification of
perceptua. Furthermore, the shaman's withdrawal of attention from
the outer world seems often to be achieved by the intermediary step of
focusing the entire attention on a highly present object to the exclusion
of everything else, just as it is in hypnosis with the focus on the hypnotist's
voice. The shaman's personal equipment (in which I include not only
actual objects and their ritual uses but also mental images and sensations
acquired by training) contains one or more element that functions as the
equivalent of the hypnotist's voice: that is, it is an object towards
which the shaman has built up the mental habit of exclusive attention.
It triggers the characteristic state of mind of the shaman during the
I have found that, when questioned about what happens during trances and
rituals, shamans emphasise seeing - meaning inner seeing. This is consistent
with the observation that where attention is withdrawn from the external
world, brain areas normally occupied in processing sensory information
begin to present experience on the basis of random fluctuations and feedback
within the sensory system. For visual centres this tends to produce a
raw material of symmetrical and geometrical shapes, which are then interpreted
as substitute visual impressions of things that they resemble, with their
appropriate emotional and contextual connotation filling in the image,
fleshing out, so to speak, the geometrical bones. At the same time, the
shaman typically dances and drums, so that the visual information is dynamic.
Physical movement dynamises and shapes the fluctuations in the sensory
systems. Hence images appear and disappear, move, approach, lead away,
fly, and so on. The state of mind of the shaman might be compared to that
of a person manoeuvring a canoe down a fast-moving stream: the difference
is that the stream is now inside the person and not in the outside world.
My conclusion is that shamans are not just actors. They do not maintain
the continuous narrative self that an actor maintains when acting a role.
In a particularly revealing interview with a Tuvan actor specialising
in playing the part of shaman in touring theatre performances, the actor
described how he was sometimes mistaken for a real shaman and invited
to heal people: the reason he did not do so was that "He did not
This is underscored by the fact that sometimes even real shamans fail
to see. In Friedholm Brückner's documentary film Boo Nar on
the shamans of Mongolia, at least one of the trances is abandoned quite
early on as the shaman decides that it is not going to work on that occasion;
this despite all the preparations having been correctly made, an audience
assembled, and so on.
Evidently religion and ritual have long been identified as distinct objects
or fields of academic study. The types of explanation or analysis offered
for these objects have tended nevertheless to see them either as results
of the general social structure and social process or as the cause of
effects required by that structure and process. Whatever can't be
explained this way is allocated either to the transcendent itself, for
those who "believe", or to psychology, for those who don't.
In the case of shamanism studies, the political history of this territorial
division is particularly evident. Thus the terms "ecstasy" and
"trance" were applied early on and reflected a Christian horror
of illegitimate and pathological forms of transcendence. (Ironically enough,
by divorcing shamanic practice from its social background this later made
shamanism highly exportable to post-Christian western societies.) Furthermore,
anyone involved in shamanism studies still has to reckon with the enduring
charisma of Mircea Eliade and his fascistic idea of a transcendent cosmic
imperative: this alone provides a strong incentive to explicate religious
experience exclusively in terms of social structure and social meaning.
This is the background against which we must understand Hamayon's
assertion (1993) that:
According to the symbolic representations of shamanic societies, the shaman's
ritual behaviour is the mode of his direct contact with his spirits; hence
it is functional behaviour that follows a prescribed pattern.
My answer to this is that if the shaman's ritual behaviour is the
mode of contact with the spirits, then ritual behaviour must be understood
in the broad sense of everything that happens to, or is done by, the shaman.
In this case the shaman whose behaviour literally and exclusively follows
a prescribed pattern is either doing a small ritual which does not require
a trance as such, or is not a very good shaman. There may well be prescribed
patterns which the shaman learns during training, but in an actual trance
the shaman will mentally grapple with spirits with their own highly unpredictable
behaviours. It is not that symbolic exchange with the spirits does not
take place, but that the transactions, negotiations, and dialogues with
the spirits are open, left open by the rituals, and that their openness
is precisely why they take place at all. This, in turn, is why these exchanges
must be represented specifically by symbols and groups of symbols, that
is to say, by using the particular open relation between signifier and
signified that we find in symbols, to mediate between indeterminacy and
Although ritual as we know it in ethnography may be rendered obsolete
by certain types of historical change, no unifying historical project,
such as that of socialism or of market-based democracy, can substitute
itself for the discontinuity of cultural phase structures responding to
the objective demands of human complexity. Cultures simply cannot be considered
as continuous entities. Nor can their special phases be considered as
functional on the level of the culture itself, only on the level of the
total context inhabited by the human being. Considered sociologically,
therefore, ritual is social only in that it arises out of the problematic
of a social being, but it does not express a given social logic, only
how that logic engages with what is intractable to it.
1. For reasons of brevity, this article is absolutely not exhaustive in
terms of covering even the main headings under which shamanism is normally
considered: in particular I have had to refrain from situating my analysis
in relation to other analyses in the literature. In a general way I have
drawn on my own field notes and recordings made in Siberia during several
extremely informal study trips made since 1990.
2. The idea of linking shamanism to hypnosis is absolutely not original,
the classic version being the adaptation made of Shor's work by Siikala
Henri Atlan Information et Auto-organisation; Organisation du Vivant,
Henri Atlan A Tort et à Raison Seuil 1986
Friedholm Brückner Boo Nar (The Messagers of the Spirits) The Shamans
of Mongolia, documentary film, 1998.
David Concar You are feeling very, very Sleepy, N.Scientist vol159, No
Daniel.C. Dennett Consciousness Explained, Penguin 1993
Gilbert Durand L'Homme Religieux et ses Symboles in Traité
d'Anthropologie du Sacré vol 1, Ed Julien Ries, Desclée,
Gedit 1992, Paris.
John Gruzelier A working model of the Neurophysiology of Hypnosis; a Review
of Evidence in Contemporary Hypnosis, vol15,p5 1998.
Roberte Hamayon La Chasse à l'Ame - Société
d'ethnologie, Nanterre, 1990.
Roberte Hamayon Are "Trance, "Ecstasy" and Similar Concepts
Appropriate to the Study of Shamanism?, SHAMAN vol 1, no 2, autumn 1993,
Molan and Kelemen, Szeged, Hungary.
Tim Hodgkinson Improvised Music and Siberian Shamanism, Musicworks No
66, Fall 1996.
Mihaly Hoppal Nature Worship in Siberian Shamanism, Haldjas Folklore vol
Robin Horton Patterns of Thought in Africa and the West, Cambridge UP
Leonid Lar The Training of a Siberian Shaman
Triinn Ojamaa The Shaman as the Zoomorphic Human, Haldjas Folklore vol
R.E. Shor Hypnosis and the Concept of the Generalised Reality Orientation,
American Journal of Psychotherapy, 1958, no 13, Lancaster.
A.L Siikala The Rite Technique of the Siberian Shaman, Helsinki, 1978
(Folklore Fellows Communications XCIII, 220).
Iannis Xenakis Towards a Philosophy of Music in Formalised Music, Indiana