Contents

Preface to the Ame ric an Edition x1

Acknowledgments xv
Introduction xv 11
PART O 1£
1. Trance and Possession 3
Tr ance 3
Trance or Ecstasy? 3 Manifestations of Trance: H ys teria and
Madness 12 The Repres entations of Trance: Shamanism and
Possession 17 Possession, Inspirati on , Communion: Identificatory
Tranc e 25 Di vinator y Trance and Mediumship 28
T he Dynamics of Possession 29
Possession Cults and Reprobate Possession 29 Music and Possession :
An Approach to the Problem 31 Possession and
Vocation 34 Trance, Crisis, Fit, "Wild"
Possession 38 Possession and O bsession 44
Initiator y T ranee and Music 46
Initiation, Possession, and Dispossession in the Brazilian Candomble and
the African Orish a Cult 47 H ebetude, Initiator y Dispossession, and
Music in the Vodun Cult in Benin 50 T he Muscial Repertoire of
Initiation 58
2. Music and Possession 63
T he Word "Music" 63
When Does Music Act' 65
What Music? 73
~ocal or Instrum ental? 73 Rhythmic, Dynami c,
Melodic? 78 Ever yda y Language 94
What Does Thi s Music Mean' 96
Who Makes the Music and in Wha t State Is He? 102
Uttering Cries 111
Musici ans and the Possessed 112
Music and Dance 114
Music, Gods, Illness 119
3 . Music, Shamanism, Mediumship, Exorcism 125
Music and Sham anism 125
Music and Med iumship 133
v1u CONT ENTS

Th e Case of the Bushmen 139

Exo rcism? 14 7
Possession and Pact Making: African Exampl es 148 Sau/and
David 154
T he Case of Tarantism 15 8
4 . The Strange M echanism 167
Rousseau and the Physical Powe r of Sounds 167 Percussion and
Drums 169 A Neurophysiological Th eory of the Effects of
Drumming 172 From H ypnosis to Conditioned Reflex and from
Emotions to Drugs 176
PART T WO
5. Music and Trance among the Greeks 18 7
Mania and Its Term inolog y \188 Telest ic Mania (Phaedrus
244d - e) 192 Music and T rance : Plato's The ory 201 Dancing
and Identifying Oneself with the Gods 206 Plato and Possession
Music; the Aulos 213 Aristotle, the Eth os of Modes, and the
Phryg ian Mode 220
6 . The Renaissanc e and Opera 22 7
From Mars yas ro Claude Le Jeune; or, T he Effects of Music 227
A Lette r on Opera 241
7. Music and Trance among the Arabs ) 255
Religious T ranee 25 5 \, ____
T he Word Samac 255 The Word Wajd 258 T he Word
Dhikr 262 Sama c Music according to Ghazzali 265
The Practice of Dhik r 271 Fakirist Practices and
Possessio n 273 The Jinn and "Poetic Furor" 279
Profane Trance 281
The \¥lord Tarab 281
T he Engendering of T ranee 283
Indu ced Tranc e vs. Conducted Trance 286 Communion Trance:
Between Possession and Shamanism 288 An Aside on
Christianism 291
Music as Emotion or as Excitation 292
Arab Musicologists and the Effects of Music 293 Sama c and
Em otional T rance 296 Dhikr and Ex citational
Tran ce 300 Willed Trance, Tru e Trance? 307 Fakirism and
Identificatory Music 309
8. Conclu sion 31 5
Music in Its Relations ro Emotional , Communial, and Shamanic
Trances 315
Music in Its Relations ro Possession T ranee 321
Notes 327
Bibliograph y 355
Conren~ ~

Discography 369
Filmography 371
1. Subject Index 373
2. Index of Religions, Sects, Divinities, and Religious Figures 384
3. Index of Ethnic Groups and Places 387
4 . Index of Authors, Individuals, an d Characters 390

Th e f igures occur on pages 251-5 4

 Preface to the
American Edition

What cultured Westerners, anthropo logists included, most commonly

think abo ut the effects of music on trance sometimes reminds me of an old
sto ry.
Some fift y years ago, the son of a great Kakemba rain maker, first of his
tribe to visit England , was taken co a football match wh ich made a great
impression on him. Back in h is village, he cold his father abou t it: "There
was an enormous crowd. You would ha ve thought it was a king 's funeral
o r something . People were seated in tight rows, around a huge empt y plot
of grass. After a wh ile , a troop of twenty men or more came running in.
Thev sca ttered and stood there for a m o ment. One of them put a round
ball down on the earth, right in the center of the green . Suddenly someone
blew a whistle. A chap rushed forwa rd and kicked the ball away with his
foot. Immed iately the rain started co pour down." His father replied:
"These big white rain makers! Their medicine must be very strong!"
ow what abou t those black drummers who can m ake people fall
down and faint on the spot? Their mus ic must be ve rv strong coo! Both are
cases of what could be called an ethnolog y of the cause-and -effect rela -
tionship.
If by virtue of its topic this book lead us into the very heart of ethno -
logy, it does so fo r another re ason as well. It happened to be written in
French by a nati ve of Paris. But most o f the data gathered in the following
pages concern othe r ethnic groups belonging to other pans of the world,
speaking other languages that express other systems of thought . Now that
this work is presented in English and meant to be read by na t ives (not only
th ose of the British Isles ) whom the aut h or did not original ly have in mind,
one m ay ask whether the reader will get the right idea of b o th the ethnolo -
gical facts cited and the theor y proposed to interpret them. A question of
translation? Of course! But transl a tion is par excellenc e a matter of ethno -
logy.
I will not linger on the ethnographic data that constitute the raw materi -
a l of t hi s book. T he y raise the ordinary problem of compilation. Suffice it
to say that the ethnographers' texts ha ve always been stric tly respected
and quoted as fully as it seemed necessary. Of course, whether the authors
correc tl y rep orted the n at ive thoughts or not still remains to be seen . This
Xll PREFACE

is the usual - and central - problem of ethnology (others would say of

cognitive anthropology, but this does nor matter ). What I will linger on is
my interpretation of those data, such as it is now presented to the reader,
in English and translated from the French.
However cosmopolitan one may be, one is always to some degree a na-
tive (if not on occasion, even nowada ys, an aboriginal ). I do not escape the
rule. The only difference between me, the author of this book, and all the
anonymous people whose notions and behavior are its very substance, is
that, contrary to the latter, who were not asked their opinion about the way
their words and gestures were reported by foreign observers, I was given
the chance of checking what had become of my writing once it was tr ans -
lated into a foreign language and of defending m y own train of thought.
Since each language is a particular organ ization of meaning, translating
a text implies a change of organization. T he relative flexibility of the latter
allows the change to be more or less important depending on the type of
translation required. In the present case, I wanted it to be as slight as pos -
sible - which of course was not always obtainable. Revising the tr ansla -
tion first made of this book, Brunhilde Biebu yck and I tried to keep the
English as close as possible to my personal way of chinking, a way which
is, no/ens volens, a French one. To achieve an acceptable result, we had to
work in close cooperation since it was necessar y to adjust nor only the ex -
press ion to the thought bur sometimes also the thought to the expression.
aturally, this procedure may have resulted at rimes in a rather marked
literalness. I hope the reader will be ready to ac cept it. I would like to add
that if in a number of places a French phrasing was preferred to an Eng-
lish, in return the final formulation chosen for the American ve rsion quite
often proved to be better than the original one and induced me to revise it.
If some day there is a new French edition of this book, I do hope it will take
advantage of these impro vements. My lifelong faithfulness to the practice
of ethnology would find in this its best reward.
No one could have carried ou t the exercise of gobetween required by
this revision with more competence, more talent and more gen tleness than
Brunhilde Biebuyck. I am glad to express all my gratitude co her.

Because of editorial constraints, the preface Michel Leiris kindl y wrote

for the original edition of this book has not been reproduced here. I regret
it deeply, not only bec ause the presence of a text by Leiris at the beginning
of this wo rk did me great honor, bur also because his views on trance and
on music would have interested the re ade r all the most since they are those
of an author well known both for his studies on possession - among the
very first ever made - and his literary writings.
For this American edition, I have brought cenains corrections to the
original French version . A few additions ha ve been made as well. But the
Preface Xlll

onl y change of some importance concerns the chapter on Greek antiquity,

wh ich has been so me what augmented and re shaped. Photographs have
also been added.

G. R.
Paris, Decem ber 19 84
 Acknowledgments

In 1968, Roger Bastide and Jean Rouch, at whose instigation the Centre
National de la Recherche Scientifique organized in Paris an international
conference on posses sion, asked me to present a paper on music. It was in-
conceivable, in their view, that such a conference could take place without
at least one session on music ; this would have meant that an essential as-
pect of the facts would be ignored . Since no one better qualified than I of -
fered to shoulder the task, I agreed to participate . However, it took all the
friendl y persuasion they cou ld muster to make me do so, since at that time
my knowl edge of the subject was confined to the very limited field obser -
vations presented on page xviii. The sheer scope of the problem appeared
formidable to me. Imprudently, but because it seemed to be the fundamen-
tal question, I entitled my paper "M usic in Its Psycho ph ysiological Rela-
tions to Possession ." In June 19 72, wh en Roger Bastide sent me the type-
script of my talk, with a request to revise it (otherwise it would have been
published as it srood ), its inadequacies seemed so glaring that I committed
the second imprudence of deciding co rework it entirely. This was how I
gradually embarked on the writing of this book. My great regret will al-
ways be that I was unable to show Roger Bastide the first draft before he
died. We owe him the most perceptive observations yet penned on the re-
lations bet ween music and possessi on . His criticisms would undoubtedly
have been inva luable to me. I neverth eless dare believe he wo uld have ap-
proved of the general direction of the pages th at follow since they tend co
show how music contributes to the socialization of the trance state, a pro -
cess whose great importance he himself had pointed out. May this book
be a modest tribute co his wo rk .
The encouragement I recei ved from my friends whil e I was working on
this book has been of the greatest value to me. My thanks go first of all to
Michel Leiris , Eric de Dampierre, and Pierre Smith, all of whom were kind
enough to read a first draft and greet it favorably. Leiris subsequently
maintained an unfailing interest in my enterprise and comforted me con-
stantl y with his advice and friendship for which I am deepl y grateful. It
was at Dampierre's suggestion that I greatl y enlarged the chapter on the
Greeks, a task I could not have accomplished without his help and that of
Jacques Lacarriere, as well as the valuable advice of Samuel Baud-Bov y.
XVI ACKNOWLEDGMENTS

Jean-Pierre Vernant was lat er kind enough to re ad this section and offer a
number of observations that I hop e to ha ve tak en sufficiently into ac-
count. Ste lla Georgoudi was kind enough to check my transliterations of
the Greek. M y heartfelt th ank s to all of them. As for the chapter dealing
w ith the Arabs, since I am even less of an Arab ist than a Helleni st , it
wo uld have been quite imp oss ible for me to write it wit hou t the friendl y
help of Bruno H al ff and Scheherezade Qas sim H assan . I thank them both
from my hea rt , as I do to the late R. P. Nyw ia fo r the comments he sent me
after h avi ng kindly consented to look ove r my text.
Among all th ose to w hom I owe helpful comments, information, criti -
cism, or support -a group too large to name- I would like to express par-
ticu lar gr atitud e to Magu y Andra !, Elena Cassin, Gisele Binon , Laurence
Delab y, Jean Gouill ard , Robene H ama yon , Georges Lapassade, Frarn;:ois
Lesure , Ar iane Macdonald, Jea n-Bert rand Ponta lis, Jean Pouillo n , Jean
Rouch , Pierre Verger , and Andras Zempleni . Various colle agues and
friend s in the Dep artment of Ethn omusicol 0gy at the Musee de ]'H omme
ha ve also earned my grat itu de for the help the y have ofte n given me dur ing
m y research, as h as the Librar y of the Musee de ]'Homme for the reading
facilit ies alw ays so kindl y an d freely put at my disposal. M y thanks also go
to Lucile Garma for her attentive and critical peru sal of th is work at each
of its many stages. Lastl y, I wou ld like to say how much I owe to Je ann e
Gousse for the friendl y and tireless kindness w ith wh ich she assisted me in
a task that, w ithout her, could ne ver have come to term.
Pierre ora also pl ayed a decisive role in the publication of this boo k in
Fr ance . To him I feel a most p articular debt of gra titude.
 Introduction

The trance sta te , a phenomenon observed throu ghout the world, is associ -
ated most of the time with music. Why? That is the question thi s book will
attempt to answer .
Why? people w ill no doubt ask. Quite simpl y, because it is music that
throws people into a trance! There is no need to look any further. ow
what if things we re not in fact quite so simple? Looking at them a bit more
closel y is enough to show that the relations between music and tr ance
could not be more changeable or more contr adictor y. In one place music
triggers it; in another on the contrar y, it calms it. Sometimes th underous
drums send the subject into tr ance; at ochers, it is the very faint susurration
of a r attle . Among cert ain people a musical instrument is said to produce
chis effect, wh ereas among ot her s it is the human voice. Some subjects go
into trance while dancing, others while lying prone on a bed. Under these
conditions, one asks oneself, How does music ace? Plato attributed the
trances of the Cor ybante s to the effects of the aulos, Aristotle to those of
the Phr ygia n mode . From the very beginning the facts have given rise to
the most opposed the orie~. To use Jean-J acques Rousse au's wo rds , is it a
physical effect or a mor al one that is at wo rk? In the twelfth century,
among the Arabs, Gha zzali upheld both possibil ities , and as proof of mu -
sic's ph ysical effects he advanced the case of camels driven into a frenzy by
che cameleer 's singing . During the Renaissance, the poets and musicians of
the Pleiade thought that music was capab le of its greatest effects onl y
through its union with poetr y. Rousseau , who constantl y took a great in-
terest in the problem, hesitated a long time befor e categorically siding
aga inst the theor y of "the ph ysical power of sounds."
A few yea rs ago, Andrew Neher, an American neuroph ysiologist ,
pro ved- or rather thought he had pro ved- that, on the contrary, the
"myster y" of the effects of drum s on trance was due simpl y to the purel y
neurophysiological action of the sounds of chat instrument. A fair number
of ethnologists and ethnomusicologists at present accept this theor y as a
given. Twenty years earlier, Melville J. Hersko vits advanced another very
different thesis explaining the effects of mu sic on tr ance as the result of
conditioned refle x. Some years later, Roger Bastide took up this last the-
ory again, but added to it the noti on of an over all situation, without which
XV!ll INTRODUCTION

the reflex in question wo uld not come into play. If we now leave posses-
sion, wh ich was Hersk ov irs's and Basride's concern, to turn to shamanic
trance and Shirokogoroff's reknowned description of it, we find what
could be called, ultimatel y, a theor y of the emotional power of music . In
yet another and totall y different view on the effects of music on trance,
Alain Danielou, an "ethnomusicolog ist of wo rld w ide reputat ion," wro te
quite seri ously in a recent number of T he Unesco Courier that "in all re-
gions of the w orld the rh ythms emplo yed [to produce trance states ) are al -
ways odd, in 5, 7, and 11 time. " And he added , unab as hed by the contrary
evidence of a thousand examples , "squ are rh ythms in four or eight time
ha ve no hypnotic effect ."
As we h ave just seen, we do not lack explanat ions. Ap art from D an ie-
lou 's , w hi ch is pure fa ntas y, an d Neher's, which is false, each has some
truth to it ; yet none is completely satisfactor y. The truth is th at the facts
are so varied and so complex they evade any single explana ti on. In a gen -
eral way, a ll-or almost all -tho se who ha ve dealt w ith trance believe,
more or less unc ons ciousl y, th at m usic has th e sec ret power to trigger it ,
and that it does so all by itself , by the interpl ay of its ow n propert ies. The y
then go o n to call up on either the power of the rh ythm , wh ich is the most
frequent a rgument , o r the magical effect produced by chants, as in the
case of th at o the rw ise ad mir able scholar, Jeanm aire, who, w hen speak ing
of th e "possessed person thrown into a trance by the demonic call of in -
cantato ry melod v," attr ibutes to music qual ities which it by n o me ans has .
Th ese are recei ved ide as th at do not stand up to ex amination.
Dem ystif ying the conception, too often adop te d, of the role played by
music in inducin g tr ance states w ill be one of thi s book's aims . The impor -
t ance of music will not be diminished for as much; quite th e contrar y. Mu -
sic w ill ultimatel y appea r as the principal mean s of manipulating the
tr ance state, but by "soc ializing" much more than by tri ggerin g it. This
proc~ss of so ci al iza tion ine vitab ly va ries from o ne ocie tY to a nother , and
t a ke s place in very different ways accord ing to the systems of representa -
tions - o r, if one prefers, the ideol og ical systems - within wh ich tr ance oc-
curs. In each case a different logic determine s the rel ati o nship between
trance and music. It is this logic chat I shal l en · to elucidate .
Although tr ance is a quasi -universal phen o menon, it is nonetheles s a
practice much more w idespread among chose peoples co nscirntin g the
ve ry object of ethno log ical research than among ocher s . Thi s bo ok is
therefo re a book of ethnolo gy, o r, more precisel y, of ethnomusicolog y.
Conseq uentl y there w ill be al most no mention of tr ance state s obse rved in
the modern Western world, wh et her among Europe an o r American Chr is-
tian sect s, de votees of pop music , or bioenerg y adepts. ot chat I reg ard
su ch phenomena as less int erestin g, but a line had to be dr aw n some -
whe re. In additi o n , especiall y in the last cwo cases, information is rathe r
sca rce.
Introduction X IX

This book had its origins in m y long-term inte rest in the music of pos -
session cults practiced in the People 's Republic o f Benin (forme rl y D aho -
me y), a stronghold of trance, the country that gave birth not onl y to H ai-
tian voo doo but to a great man y other Afro-Cuban and Afro-Brazil ian
cults . Ten or so period s of va rying length spent in that country allowed m e
to attend - usuall y in order to rec o rd music - a fair number of ceremonies
directly connected w ith pos session amon g the Gun, the Fon , and the Yor-
uba. But m y first encounter with the trance state d ates to a da y in 1946,
durin g a vis it to the Congo, when right before my eyes a Babinga Pygm y
was suddenl y entranced in the ver y m ids t of a hunting d ance, whose mus ic
my companions and I w ere recording. Two years later, in Bah ia , Pier re
Verger introduced me to the candomble of N uestr a Senhora on the occa-
sion of a beautiful nighttime ceremon y for the wa ter of Ob ata la, which
wa s followed on the nex t da y by a feas t during which a number of posses -
sion s occurred. On va rio us occasions I have been fortunate enough to at-
tend o ther such ritu als elsewhere: in Dak ar, in particul ar, w here in 1952 I
was a wide -eyed spectator at a ndbp ceremony w hich I sh all never forget,
and seve ral days later a w itn ess - and one who was somewhat too roughl y
handled for my taste - to a "devils " ceremo ny among rhe Sarakole . Twenty
years later , in Niame y, hav ing pro vide nti ally mi ssed my plane, I wa s able
to attend, in rhe comp any of Je an Rouc h this time, the ritual enthrone -
ment of a holey daughter am ong the Songhay. Ho wever supe rficial they
ma y have been , the se contacts wit h ocher aspec ts of possess ion ha ve been
extremely inst ructi ve for me .
In recent years, and with th is book in view, I have o rgan ized my re ad -
ing s around these direct obse rva tion s of rhe rel ation s between music and
trance ; the y ha ve progressi vely ex tended themsel ves to o cher region s of
Africa, then the Mediterranean world, Asia, Indo ne sia, and from there ,
through sha manism, to Indi an America. Quite arbitr aril y, and in order to
limit the already broad scope of m y re sea rch , I have excluded noron ly
China and Japan bur also Polynesia an d Melanesia. I must also add that
even w ithin the bound ar ies just indic ated, I do nor in any way pretend to
hav e exhausted rhe available liter ature, w hich I have onl y gleaned. Im-
mense lacunae remain. Neve rthe less, I do feel that the sa mple cases I pre -
sent he re are repre sentat ive enough and th at the different types of rela -
tions th at emerge from them are to be seen as a va lid basis for further
research. If thi s proves to be the case, this book w ill then have ampl y ful -
filled its aim.
I hope I ma y be permitted o ne last observation. This wo rk co vers too
man y diffe rent domains not to be vuln erable - abundantly, I fear - to
critic ism from the spec ialists concerned. It is the overall picture that is im -
port ant, howe ver. If it finds favor in their eyes , then I trust the y w ill con-
sider that the weaknesses of det ail are the ine vitable price to be paid for
such an enterprise, and th at the y should be w eighed ac cordingl y.
 PART ONE

As long as we choose to consider sounds onl y th rough the co mm otion the y st ir in

our ner ves, we w ill never ha ve the tru e principles of music and o f its power over
o ur he a rts.
J.-J.Rousseau, Essay on the Origin of Languages
1 Trance and Possession

TRA 1CE

Axiomaticall y, trance w ill be considered in this book as a state of con -

sciousness composed of two components , one psychoph ysiological, rhe
other cultural. The universalit y of tr ance ' indicate s char it corresponds to
a ps ychoph ysiologi ....al disposition innate in human nature , although, of
course, de veloped to va rying degree s in different individuals. The vari-
ability of its manifestations is the result of the va riet y of cultures by w hich
it is conditioned.
A scare of consciousness , trance cons ist s, for the subject, in a particular
experience composed of a series of events char can on ly be described by
chos e who h ave lived through them. Although it constitutes the very es-
sence of the phenomen on, for all sons of reasons I sha ll nor make use of
such subjec ti ve experience in order to delimit the concept of trance. In-
stead, I shall define it by referring to its external m ani festations, the con-
text in w hich th ey can be obse rved , and the represent at ion s of whic h they
are the object. Howe ver, those who ha ve dealt thus far with trance have
not alw ays used the same vocabular y to identif y the same manifest at ions,
assoc iated with simi lar contexts and representations among different pop -
ulations. Sometimes, in fact, the ver y same autho r uses two or three differ -
ent terms to refer to one and the same phenomenon. Ir is importa nt at this
stage to unify our terminolog y, since ot her w ise we will nor know what we
are t alking about. T he first point to be clarifi ed rel ates to the use of the
wo rds "ecstas y" (extase ) and "trance" (transe ). I shall de al w ith the word
"crisis " o r "fir" (crise ) later.

T ranee o r Ec stasy?
In French, th e wo rd transe is often used in ethnological literature as a syn -
onym fo r extase. In ever yda y usage, howe ver, the rwo words signify ver y
different things. The same holds tru e in Engl ish, w ith the one exception
th at w hereas "ecstas y" and "trance " stand in more or less rhe same oppos i-
tion in medical and colloqu ial us age, their usage is the reverse of the
French. Two quotations, both from medical diction ari es, make th is clear.
The English work 2 defines tr ance as "a sleeplike st are, as in deep hypnosis,
r
I 4 CHAPTER ONE

appearing also in hysteria and in some spiritualist mediums, with limited

sensory and motor cont act, with subsequent amnesia of wha t has oc-
curred during the stare." The French work' defines extase as "a mental ·
stare characterized by profound contemplation ac companied by a loss of
sensitivit y and morricir y." Thus one calls transe what the other calls ex-
tase. Let us no w shift from medicine to social anthropolog y, beginning
with English. The w ord "ecstasy" does nor appear at all in the index of
Jane Belo's work, Trance in Bali (1960 ). Both "ecstasy" and "trance ," on
the other hand, occur in I. M. Lewis's Ecstatic Religions ( 1971 ), which be-
gins with a chapter entitled "Towards a Sociology of Ecstasy." One would
expect the author to define ecstasy, but he onl y proposes a definition of
"trance " (taken from the Penguin Dictionar y of Psychology, and very close
indeed to that from T abe r quoted here above ), a term that, as he says, he
uses in "its general medical sense" ( 1971, 3 8 ). In the course of his book he
certainly uses "trance" most frequently, but this does not prevent him from
using "trance" and "ecstasy" as if they we re absolutely synonymous. 4 In
the vo lume of collected papers edited by Beattie and Middleton under the
title of Spirit Mediumship and Society in Af rica (1969 ), one might search
in vain for the wo rd "ecstasy," which does nor appea r in the index either:
"trance" is used throughout. On the other hand, the Arabic word wajd,
wh ich among the Sufi denotes a state that both British an d French ethnolo-
gists wo uld indisput ab ly render roday by the word "trance" or transe, was
invariably translated as "ecstasy" by Macdonald in his 1901 translation of
Ghazzali 's well -known book .' In his wo rk on Sufi seers, Trimingham
( 1971 ) remains faithful to this tradition throughout, w ith one very signifi-
cant exception. When he needs to describe precisely what the wajd con-
sists of in practice (1971, 200 ), he introduces the word "trance," and uses
the terms "e cstatic trance " or "ecstasy or trance-like stare." Lastly, Erik a
Bourguignon (1965, 1968 , 1973, 1976 ), to whose works on possession I
shall often refer , constantl y uses the word "tr ance. " On rare occas ions she
also uses "ecstasy" in order ro denote rhe parox vsm of trance. 6
We can thus see that the me aning assigned to the words "trance" and
"ecstas y" poses a ver y real problem. Finall y, I should point out that neither
"tranc e" nor "ecstasy" figures in the general article on possession pub-
lished in the Enc yclopaedia of R eligions and Ethics o f 1918 (the first of it s
kind '). There are sever al reference s to "ne rvo us crisis." "Ecstas y," on the
other h a nd , does appear in the articles dealing wit h possession among
Greeks, Muslims, and Christians.
In French the situation is just as complic ated. W hen dealing with pos-
session among the Ethiopi ans of Gond ar ( 19 5 8 ), Michel Le iris more often
th an nor uses the word crise (crisis ), but he sometimes also uses transe,
particularly with reference to dancing, a "classic means of inducing
trance " (1958, 18 ), an d oc cas ionall y speaks of "ecs tatic techniques"
Trance and Possession 5

(ibid. , 93 ). The same terminolog y prevails in the wo rk of Andras Zem-

pleni ( 1966 ). When describing posse ssion ceremonies among the Wolof in
Dak ar, he generally uses crise, like Leiris, but in the context of dancing, he
also makes use of transe (1966, 408 ) and refers elsewhere to la transe ex-
tatiqu e (ibid., 434 ). In her contribution to rese arch on the H ausa bori cult
in Niger , Jacqueline Monfouga -Nicolas (1972, 177 ff. ), on the contrary,
constantly uses the word transe, never uses ex tase at all, and scarcely ever
uses crise except to designate a very specific case of profane possession in-
duced by drugs (ibid . , 198-99 ). To provide a final example, not itself Af-
rican but at leas e of African origin, when describing the Brazilian candom -
ble in Bahi a , Roger Bascide (1958 ) uses transe and extase as practically
synonymous. Moreover , although his chapter de voted to this aspec t of the
cult is entitled "La structure de l'extase " (ibid., 17 3 ff. ), the same phenom -
ena are examined under the title "La tr anse" in a lacer collection of articles
(1972 ) entitled Le reve, la transe et la Jol ie. le is true rh ac rhe four cults I
ha ve just mentioned , the candomble in Brazil , rhe bori cult in iger , as
we ll as rhe rab culr in Senegal and rhe zar culr in Ethi opia, do displa y acer -
t ain number of differences. Bue all four are cults o f possess io n and , except
for the degree o f violence involve d, th ey are comparable nor only in their
manifestatio ns but also in the conditions under which the y occur and in
rhe ideas asso ciated wi th chem. As we ha ve seen, what o ne writer calls a
"cr isis" th e other terms "trance " and the third "ecsta sy."
Lee us no w mo ve o n to cwo eq ually ren ow n wo rk s, both very different
from chose already mentioned: Jeanm ai re 's Dion ysos and Mircea Eliade 's
Le Chamanism e, both of which appeared in 1951. Althou gh the second of
the cwo is sub titl ed "Le s technique s archaique s de l'excase " (Archaic tech-
niques of ecstasy ), the word transe is ne vertheless employed almost as of -
ten as extase .; We re ad , for example , chat among the Samo yeds "g reat
sha man s undertake the ecst ati c journey in a trance. " Jeanm aire , for his
pare , use s transe much more frequentl y th an extase, but when referring to
rhe dich yramb, for example, he speaks of "col lecti ve ecstasy" ( 19 51 , 248 ),
a nd often resorts to rh e adjecti ve extat iqu e, p articularl y in the phrase
trans e extatiqu e (ibid., 252 ). The sa me use of extatiqu e can be found in
Luc de Heusch (19 71 ), in w hich w e find "ecst atic religions ," "ecstatic cri-
sis," a nd "ecst atic cheater " side by side w irh rhe wo rd transe on irs own,
which he uses fa r more frequentl y th an ex tase . {It sho uld be noted that the
frequent use of exta tique by so man y author s can clearl y be at tributed to
rhe fact that this word is an adjective. If there w ere an adjecti va l form of
transe, the a uthors I h ave quoted would in man y ca ses undoubtedl y ha ve
p referred it. ) One final wo rd on rhe interchangeabilit y o f ex tase and
trans e: in the collection Les danses sacrees, published in 1963 by a group
of ori entalists , numerous p assag es are specificall y de vo ted to tr ance; un-
de r che term transe in the index, how ever, one is referre d ro extase . 8
6 CHAPTER ONE

This brief summary of the use of the words transe l trance and extase l
ecstasy in modern ethnology clearly demonstrates that their usage is very .
inconsistent. 9 The two wo rds are never used to establish an oppos ition be-
tween two different states. Yet, as we shall see, this opposition does exist,
and it is essential to convey it. Within the perspective of this book, it is
even more indispensible to make this distinction since trance and ecstasy
each stand in very different relationships to music. We w ill, therefor e,
have to establish the distinction between them clearly, so th at we can put
an end to the confusion and agree upon exactly what we will me an from
now on by the words trance and ecstas y.
Everyo ne will admit that in French it is the wo rd extase that springs to
mind whe n we think of the mystical experience described by Saint Teresa
of Avila, and it is this very wor d (Spanish extasis ) that she herself uses in
her wr itin gs. But she is careful to make a distinction between its tw o as -
pects: "union" (union ), which constitutes the weake r form of ecstasy, as it
we re, and "rav ishment " (arrobamiento ), which "goes far beyond union,"
"produces much greater effects,"' 0 and thus corresponds to ecstatic pleni-
tude. "Union," a frequent state with Tere sa of Avila, can be ach ieved un -
der almost an y circumstances. "Ravishment," on the other hand is linked,
as we shall see, with silence , solitude, and immobility. Before describing
her famous vision of the angel , which recurred on several occasions, T ere -
sa of Avila analyzes her state as it comes to her: "The whole body is
broken and can move neither feet nor arms; if it is standing it collapses as
though drawn downward by some great force and can scarcely breathe." ''
So much fo r imm ob ility. As for solitude and silence, we all know the y are
the rules that she herself wanted for all Carmelites. Speaking of a cloister
dear to her heart , she writes : "Th ose who seek solitud e" in order to enjoy
Chris t , their Husband, have every possibilit y of living constantly in his
compan y." ' 3
Silence, solitude, immobilit y, three conditions that represent the exact
opposite of those required by the shaman when he officiates or by the pos-
sessed person when he dances. In these tw o cases an altered state is
achieved in public, accompanied by music , and manifested , often very vio -
lentl y, either by dancing, chanting, or both - in short , by the ver y opposite
of immobilit y. I could cite such other example s as Dion ysiac dances in an -
cient Greece, the Haitian voodoo cults, the brotherhoods of the Maghreb,
such as the 'Aissaoua, Gnawa, or others, the t arantula victims in southern
Ital y, and, lastl y, the sh am anic curing ceremonies , be the y Siberian,
American Indi an , or Eskimo. All these examples involve noisy, more or
less frenzied behavior, and call for a gre ater or lesser degree of ph ysical
agitati on. There could thus be no greate r dissimilarity between these two
types of behavior, that of Ter esa of Avila on the one hand and of the sha -
man or possessed person on the other. In current French usage, the term
T ranee and Possession 7
extase rather than transe is applied to the first, and transe is more fre-
quentl y used than extase for the second.
Considering the primar y meaning of these words, this is hard to justify.
The Greek ekstasis signified first an d foremost "the act ion of moving
through space, displacement, deviation." Hence "dist urbance, agitation,
wandering as applied to the mind." 14 (One might have expected Plato to
use it in the context of mania, but he never does. In fact the wo rd did not
figure in his voca bul ary at all. " It was only fairly late that ekstasis ac-
quired the meaning it still has today in the vocabulary of mysticism. ) On
the other hand, trans e also designates a change of state, but of a different
kind, originally related to the act of transir, that is to say of passing away,
and thus of entering into a state of definitive immobility. (It should be not-
ed additionally that death - as a transition - is very often accompani ed by
agitations of greater or lesser degree, by convulsions. For this reason,
"trance" is doubly suited to a description of states of possession, since such
states often entail a convulsive stage, and in numerous cases are linked to
symbolic death. ) It was appa rentl y spiritism that first gave the wo rd
"trance" the meaning it currently has toda y in the anthropology of reli -
gion. It was used in the late nineteenth century to denote "the state of the
medium when depersonalized, as though the visiting spirit had taken him
or her over." 16 But the life of words is not governed solel y by the logic of
their etymologies. If "trance " has gradually replaced "ecstas y" in descrip-
tions of possession states, it is perhaps because , in both English and
French, the word trance in some wa y evokes both trembling and dancing ,
and so summarizes, or better yet (like Lewis Carroll ), telescopes the two
aspects that are particularl y characteristic of such states.
Whate ver the case ma y be , it is evident that despite certain hesitations
and some inconsistencies associated with using the adjective "ecstatic, "
the tendenc y toda y in both languages is to use "tr ance" to design ate, in a
general wa y, the agitated states I just mentioned, and to use "ecstac y" less
and less. Personall y, I would be in fa vor of specializing the use of the two
terms, of using "ecsta sy" solel y to describe one particular type of state - al-
tered states , let us sa y, attained in silence , immobilit y, and solitude - and
of restricting "trance" solel y to those that are obtained by means of noise,
agitation, and in the presence of others.
In order to demonstrate the distinction between ecstas y and trance I
have thus far referred to only one ex ample of ecstas y- not because it is
particularly representati ve, but because it is well known - that of Saint
Teresa of Avila. Before going further, I would like to broaden our com-
parative base by adding further examples to the "ecs tatic " category: the
state that the early Church Fathers " hoped to attain in the deserts of
Eg ypt; the one sought by certain Tibetan monks in their walled -up soli-
tude; 18 and that still pursued in our own day in Senegal by the Wolof "ma r-
8 CHAPTER ONE

abouts" who withdraw into grottoes in order to practice khalwa, and in

this way obtain divine visions or jenneer. 19 Let us also include the sta te of
annihilation in God (Jana<)sought by the Sufis when they practice a cer-
tain kind of prayer (the dhikr, said to be "of the intimate, " to which we
shall return later ), ' 0 the state of samadhi, that is to say, annihilation at-
tained by Indian yog i, and lastly the one achieved by Japanese mystics
through nembutsu. " In his comparative study of mystical practices cen-
tered on dhikr, which is to say the recolleciton of Allah by ceaseless repeti-
tion of the di vine name, Louis Garder, subsequent t0 Louis Massignon,
stresse s the shared aspec ts of Sufi and yoga experiences in their quest, the
former for the state of Jana ' ("ann ihi lat ion " in God ), the latter for samadhi
(the "abolition of all action" ). These two experiences, however, occur
w ithin general conceptions of the world that are by no means the same. To
quote Jean Filiozat's comment on samadhi, '' the "true nature of this state"
consists not in "being ravished out of oneself" but in stab ilizing "one's
thoughts and latent psychic tendencies" and to succeed thereby in dis-
carding "all representation. " The Teresian srate of ravishment ,fana<, and
samadhi can certainly all be seen as ecstasies. But the y are three very differ-
ent forms of ecstasy. Mircea Eliade ( 1948, 93 ) suggested use of the word
"enstasis"' 3 to designate the State of samadhi, precisely in order t0 distin-
guish it from ecstas y. This word has often been used since, notably by L.
Garder (1953; 205ff. ) Other d istinctions mu st still be made if we are to
seria te the various aspects of "ecstasy" in any definitive order, since the
word encompasses such a vast variet y of phenomena. Thi s is not our pur-
pose here. From the viewpoint of chis book , what we need to remember is
that enstasis and ecstas y must be seen as forming rngether a categor y that
is distinct from that of trance.
Whe ther expe rienc ed by Christi ans, Muslims , or Buddhists, wh ether
observed in Spain, Senegal, Ti bet, India, or Japan, the st ate s w ith wh ich
we are dealing here all are achieved in solitude. "Loss of ever ything by
means of radical withdrawal and aloneness," is how Garder (1952, 676 )
defines Jana', the "annihilation in God" of th e Sufis. Elsewhere (1953 ,
208 ) he writes that samadhi is achieved '\ ,·hen the accent is placed on
aloneness, solitude filled with self in self. " And what do the marabouts of
Seneg al seek in the depths of their dark grorrnes, o r the Tibet an monks in
their walled -up retreats, if nor a similar rad ical solitude' Such a solitude
therefore seems to be a nece ssa ry conditi o n for the b losso ming of or prep-
aration for ecsta sv. The contrar y is true for tr ance .. ..\nd there are still oth -
er features that distinguish the two scares. As \\'e shall see , trance usuall y
comprises a convulsi ve stage, accompanied by cries, trembling, loss of
consciousness, and falling . Nothing of chis o rder occ ur s in Jana', sa-
madhi, or in the weak form of T eresian ecst as,·. There is also another ,
Tr ance and Possession 9

even more important difference: ecsrasy is a keenly memorable experience

w hich one can recall and ponder over ar leisure and which does not give
rise to the dissociarion so ch aracte risric of trance. The great Christia n
mystics, Saint Teresa of Avila, Saint John of rhe Cross, Saint Ignatius Lo y-
ola, an d man y ochers, have abundantly commented on their ecstasies. T he
same is rrue of the Church Farhers and of Indian ascet ics. Trance, on the
contrary, be it of a pos sessed person or of a shaman, is characterized by to-
tal amnesia. The relationship of the self to the trance state is, in this re-
spect, diametrically opposed to th at of the self to ecstasy.
Oesterreich (1927, 23-24 ) percei ved this essential difference between
possession tr ance and states of ecstas y. "On the latter, " he wrote, "we pos -
sess a hose of autobiographic al sources .. .. Descripri ons of possession by
chose w ho h ave experienced it are, on the contrary, extremely rare ." And
he added: "T his scarcity of accounts by those who have undergone the ex -
perience has a profound ps ycholog ical explanation that stems from the
very nature of the state of po ssession. We are dealing in part with states
that entail more or less total post facto amnesia, so that , gene rally speak-
ing, the majority of th ose who are possessed a re in no position to describe
their condition."'" He does nevertheless cite a few exceptions-including
some ve ry noted ones: Sister Jeanne des Anges, the most famous of the
Loudun nuns, and Father Surin, one of her exorcists who in turn became
possessed 2 ' - in w hom the state of pos sessio n rem ained , to use his expres -
sio n , "lucid," me a ning th at the subje ct was conscious of being inhabited
by th e "de m on" (since he usually is concerned ) and subsequently recalled
that state. In these cases we are dealing exclusively with a t ype of posses -
sio n I shall subsequ entl y term "reprob ate" 2 • in order to distinguish it from
"cultivated" possession, the kind th at wil l primarily concern us in this
book . Thi s being so, the exception represented by this "lucid form of pos -
session" as opposed to what Oesterreich calls th P "sleepwalking form ,"
which is always followed by amnesia, does not invalidate the rule. The
latter, wh ich has been described numerous times, states that the possessed
pe rson ne ver, in principle ,2; remember s anything about his trance follow -
ing his return to consciousness. One may question the sincerit y of this for -
getfulness or wo nder abou t the mechanism at work, but this woul d
change nothing . The same ho lds true of the sha man, as is evident from
many examples including what G. Morechand says in his study of sha -
manism among the Hmo ng of Laos ( 1968, 202 ):" 1o Hmong shaman wil l
eve r admit to remembering the trance he has just experienced, or ever
agree to calmly, ' intellectu ally' repeat any pre viously experienced
trance . . . This is an absolu te rule among the Hmon g . 1o one will ever
be able to persuade a shaman to perform his chant outside of the trance it-
self. Shamans unanimously ag ree th at once the y ha ve returned to their
10 CHAPTER ONE

normal state, following the trance, the y remember nothing .... : they
have 'trembled, ' then 'woken up again,' and none of them will accept deni- ,
al of this rupture."
Among the Shakers of Saint Vincent in the Antilles - a Christian sect
whose trances belong to the type I shall subsequently term "co mmunial ,"
since there is no possession in the stri ct sense - the trance state is similarly
followed by amnesia (Henn ey 19 73, 239 ), even though it is only partial
and -signific antly enough-relates solely to the context. Still other types
of trance are accompan ied always or usually, totally or partially, by amne-
sia. Even though it is not possible to say as much, with no exceptions, for
all the forms figuring in the general table of trance drawn up at the end of
the book (seep. 289 ), amnesia characterizes the principal types of trance
strongly enough to justify using it as a distinctive trait differenti ating
trance from what I propose to term ecstasy.
Two further traits can be added to the list of oppositions already cited
w ith a view to distinguishing between ecstasy and trance. The former is
frequently accompanied by visual or auditory hallucinations. In the first
of his series of three volumes entitled Ecstasy or Religious Trance (a per -
fect example of what has been said previously about the usage of these two
words! ), Arbman reserves an important place for visions and hallucina-
tions. These are indeed characteristic of Christian ecstasies, with which he
mainly deals in the first part of his wo rk. "The ecstasies of Saint Anthony
( . .. and many others ... ) have explicitly the character of a vision, " we
read in the chapter on ecstasy among the Church Fathers in the D iction-
1wire de spiritualite (1961, col. 2104 ). Tr ance is usually free from such
hallucinations. I should also observe in passing that as far as possession
trance is concerned-the type that interests us the most here-the absence
of hallucination conforms to the very logic of the relations between the
possessed person and the divinity possessing him or her. Since the divinit y
takes the place of the possessed person's everyday personality, it does not
constitute an appar iti on for the subject, contrary to what often occurs
during ecstasy.
Finally, ecstasy and trance differ in that while trance is always asso-
ciated with a greater or lesser degree of sensory overstimulation-noises,
music, smells, agitation -ec stasy, on the contrary, is most often tied to
sensorial deprivation -si lence, fasting, darkness. othing could be more
striking in this respect than the contrast one can observe among the Wolof
of Senegal between the behavior of marabou ts who seek out ecstasy in the
silence, solitude, and darkness of their grottos and that of the practition-
ers of the ndop, who enter into trance in the midst of a dense crowd, stim-
ulated by drink , agitated by wild dancing and the din of drums.
Ecstasy and trance may therefore be characterized in relation to one an-
other by two series of terms forming the following oppos itions:
Trance and Possession 11
Ecstasy Tra nce
immobility m ove ment
sile nce noise
soli tude m co mp any
no cn sis crisis
sensory deprivation sen sory oversti mulation
rec ollec ti on amnesia
hallucinations n o hallucinations

W hen all th e characteristics in one of these n vo series are present, we

then are dealing w ith what one might term the full form of either ecstasy or
trance. Clearly, h owev er , such comp leteness is no t always rea lized, or the
configu r ation of features pro vided by a given state ma y in fact be compos-
ite. In the first case, when we are de aling wi th full forms, the difference be-
tween the two sta tes is quite clearcut, in the second much less so. Ecstas y
and trance must therefore be regarded as constituting the oppos ite poles of
a continuum, 2 8 w hich are linked by an uninterrupted series of poss ible in-
termediary states, so that it is sometimes difficult to dete rmine w hich of
the two is in volv ed. T o cite onl y one exa mple , that of Teres a of Avila
(which, as we have seen, can be justifiabl y said to pro vide the very model
of ecstas y ), in at least one case (albeit a very exceptional one, it seems ) we
are confronted with what I shall henceforth identi fy as trance. I refer to the
"ecstas y of suffer ing" that overtook her one even ing when "in the company
of the sister s . . . , we sang a number of verses about the torments we suf-
fer from living without God ." "Since I was alre ady experienc ing thi s tor-
ment," Saint Teresa wr ites (1 949, 500 ), "I was so gripped by it that my
hands stiffened despite all my efforts . .. ," and she adds that the suffe rin g,
"now so int olerable ... forced me to utter loud cries that I was powerless
to repress ." One cannot help but compare this account with other exa m-
ples mentioned in the chapter on the Arabs, involving the particula r form
of trance I w ill suggest we term "emotional. "
Finally, I would like ro obse rve that trance and ecstasy m ay both be
practiced, albe it within the context of different ritual s, by the sa me indi -
vidual within the same religious faith . Thu s th e Shakers of Saint Vince nt
in the Antilles, who still practice tr ance as I have just defined it , also atta in
ecstasy when they wi thdr aw inro the "sec ret room" (Henne y 1973, 242 -
43 ) in o rd er t0 weep awa y thei r sins. In isol at ion , immob ility, and through
corporeal chastisement, they then undertake a "spi ritu al journey" th at is
apparently strongly characterized by "hallucinator y act ivities. " 29
Now th at the difference bet ween trance an d ecstas y- and also, I h op e,
the gradatio n s exist ing benveen them - h ave been adequa tely established,
let us m ove on to th at which for us is th e essenti al point: the differences be-
nveen trance and ecstas y in their re spect ive relations to music . W herea s
12 CHAPTER ON E

tr ance, as we shall see throughout this book, is very frequently and very
closely assoc iated with music, ecstasy, as it has just been defined, never
makes use of it at all. There is an inherent incompatibility between the
practice of ecstas y and music. This is, after all, only logical. Excluding
certain deliberatel y contri ved experiences, immobility, silence, and senso-
rial deprivation are incompatible with music. Indeed, in order to escape
from her ecstasies - which we re so strong that she succumbed to them
against her will (F. Jene 1961, 2147 )- Marie de ]'Incarnation, more or
less a ~ontempor ary of Teres a of Avila, would begin playing the spinet. In
a quite different form of ecstas y, that of the yogi, it ma y happen that music
does pla y a role under the form of what Mircea Eliade (1948, 254 ) calls
"mystic sounds," which are heard during yoga meditation. The vo ice of
the divinities heard in such states is said to be "similar to the sound of a
golden bell." In other wo rds, the sounds that are heard are auditory hallu -
cinations, and thus a further confirmation of our definition of ecstasy. In
her labor atory study of yoga techniques, Therese Brosse ( 1963, 116 ) tells
us that in bhakti y oga a song is transmitted mentall y and silently from
guru to disciple as a means of enabling the latter to attain dhyana, the
higher stage of samyama, which represent s the fullest stage of ecstasy.
Here again we ma y view the music in question as hallucinatory in nature .
As we can see , this case does nor constitute a counterexample any more
than the preceding one; quite the cont rar y, in fac.r.
To pur an end to the debate o n the possible relation between music and
ecstasy, suppose someone says that a Mozart quartet plunged him into ec-
stasy at a concert. This is, of course, pure verbal inflation, or, if one pre-
fers, hyperbolic use of language. Besides, how does chis ecstasy manifest
itself? Nor by noise or frenzy, certainl y, but rather by a great desire to pro-
long it in silence. In practice , applause usuall y brings ir to an end. Ir is,
then, trance-or enthusiasm, to which we shall return later - that drives
ecstas y away. Ultimately, we could also invoke Scriabin and his Poem of
Ecstasy . This would lead us far astray.
All this having been said, I shall not return to the subject of ecstasy in
this book. The task fac ing us now is to isolate and describe all the other
signs by wh ich trance ma y be recognized.

Manifestations of T ranee: H ysteria and Madness

Trance is a temporar y stare of consciousness, or, as the word itself indi-
cates, ir is transitory. One lea ves one's usual state when one enters into
trance , and at the end of a certain period of rime, the duration of which
va ries considerabl y from case to case, one returns to one's previous state.
T ranee is thus an unusual stare . One might conceivably say of someone
that he or she lives in perpetual tr ance, bur the fact remains that such a
state is reg arded as unusual , or in this particular case as unusuall y usual.
T ranee and Possession 13
But w hat are the telltale signs of tr ance? Let us divide them into two
categories: symptoms and beha vior. I shall term symptoms those signs
that constitute merel y the simple , unelaborated expression of a certain
pertu rb ation experienced by the subject at, let us say, the animal level (be-
ing seized by tremblin g is a sign of great fear in both men and ani m als).
"You tremble , carcass!" Tur enne sa id to himself. On the other hand, the
runaway horse perhaps repre sent s th e mos t strik ing example of what we
would call "an im al trance." I shall term behavio ral those signs that no
lo nger constitute a simple re action, as do symptoms , but a po sitive action
endowed w ith symbol ic value. (It must be ack no wledge d in pass ing that
the boundary between these two cate go ries is impo~sible to draw wit h any
precisi on. Thi s difficult y, however, is inherent in any, or almost any, at -
temp t at categorization. Such categorization still rem ains ind ispensible,
howe ver, since without it the multiplicity and divers ity of such signs
w ould remain sheer ch aos. T he choice lies between inevit ably ar bitra ry
and reductive categories - which rem ain open to recons ider atio n , be it
said- and confusion .)
W ithout att empting an inventory, let me list the princip al symptom s of
the tr ance state: trembling ,3° shuddering, horrip ilation , swooning, falling
to the ground, yaw ning, lethargy, convulsions, foaming at the mouth, pro -
truding eyes, large extrusion s of the tongue, par alysis of a limb, thermal
distu r ba nces (icy hands despite tropic al heat; bein g hoc despite extreme
circumamb ient cold ), insensitivity to pain, tics , nois y breaching, fixed
scare, an d so on. In addition , there are two signs chat are difficult to cate -
go rize as either purel y symptom at ic or beh av ior al. First, the subj ect gives
the impre ssion th at he is totally eng age d in his trance, that the field of his
conscio usness has been completely taken over by chis st ate, chat he has lost
all refle x ive cons ciou snes s, that he is incapable of coming back to hi mself
(unless enjoined to it by some externa l intervent ion ); in short, th at he is
plunged int o a so rt of bewilderment . It is imp oss ible to attr act his atten-
tion; if he turns his eyes in your direction he does not see you. Seco nd
(somewhat complementary to the first ), onc e the subject has emerged from
his trance, he has no re collection of it.
As for the beha vioral signs, which are even more di verse - not surpris-
ingl y so , since in their case the pr ed ominant facto r is no longe r nat ure but
culture - ultim ately the y all seem co be merel y the various "sign ifiers " of
o ne a nd the same "s ignified ." 3 1 One could sa y that in pr actice the y alwa ys
symboli ze the intens ification of some p articul ar facult y by means of an ac-
t ion endowed w ith certain extraord inary or astonishing aspec ts . Thus
trance m ay be recognized, among othe r signs, by the fact that one can
wa lk on burning coals without being burned, pierce one 's own flesh w ith -
o ut bleed ing, bend swo rd s one would no rm ally be un a ble to cur ve, con -
front danger without flinching , handle poisonous snakes w ith out being
bitten , cure diseases, see into the future, embod y a divinit y, speak a lan-
14 CHAP TER ONE

guage one has never learned, swoon or die of emotion, be illuminated by

the Eternal, enter into contact w ith the dead, travel in the land of the
gods, confront those gods, emit totally unhuman cnes, give acrobatic dis-
pla ys beyond one's normal ab ilit y, bend backwards ro make a perfect arc,
compose poems in one's sleep, sing for days and nights on end without a
break, dance without difficulty despite being crippled. Thus trance al-
ways manifests itself in one way or another as a transcendence of one's
normal self, as a liberation resulting from the intensification of a mental
or ph ysical disposition, in short, as an exaltation - sometimes a self-
mutilating one-of the self. In the earl y days of Christianity, people in
trances we re called energoumenoi 32 since they were thought to be prompt-
ed by a supernatu r al energeia attributed, according ro circumstances, to
God, the Hol y Ghost, or the Devil. The word energume ne was still in cur-
rent use in seventeenth -centur y France to designate people who we re pos-
sessed, notabl y the nuns of Loudun. 33
These beha vioral signs can vary, needless to sa y, from the very spectacu-
lar to the extremely discreet, just as the symptoms listed earlier can vary
from the extremely visible to the almost imperceptible.
The individual in a trance state is chus rec ognizable by the face that ( 1)
he is not in his usual scare; (2 ) his relationship to the world around him is
disturbed; (3 ) he can fall pre y to certain neurophysiological disturbances;
(4 ) his abilities are increased (either in reality or ocherwise 3 ' ) ; (5 ) this in-
creased abili ty is manifested by act ions or behavior observable by others.
The picture can obviously be either more or less complete or incomplete
depending on the indi vidual case. In case of doubt, the context will indi-
cate whether or not one is justified in sa ying chat trance is invo lved .
The word "liberation " has just been used. Ir is in connection with the
quest for a liberation of this kind chat, in an article intended to situate his
work in relation (and in opposition ) to the "human potential movement"
in California, Max Pages speaks of trance: "When I conduct a seminar I
have no fear of admitt ing chat I am in a sort of secondary state akin to
trance," he writes (1973 , 171 ). This "liberation of the body" constitutes
the goal of the "new trance groups," especialh · th ose stemming from the
bioene rgy movement, also mentioned br Georges Lap assa de (1976, 215 ).
What is chere in common bet ween tr ances of rh,s kind and those I shall ex-
amine in this book? We knO\\. too little ab u ur them to decide on this point.
Accordingly, I shall not take them into account, especially since their rela-
tion to music seems ve ry uncertain . With reference to the seminars he de-
scribes, Pages (ibid., 17 0 ) notes that in the "expressive equipment at the
disposal of the p articipant s," there are "rudimentary musical instruments,
particularl y those requiring no previous tr aining that serve to produce
simple rh ythms: tom -rom s, drums, kettle drums, cymbals, xylophones,
flutes . . .. " Rudimentar y instruments, simple rh ythms . . . A strange
conception of music, indeed'
T ranee and Possession 15
Whether one likes it or nor, the signs I have listed as characteristic of
trance are far too clearly reminiscent of hyste ria " to avoid a discussion of
this matter here. 1 or is it merel y hyste r ia that should be cons idered.
T ranee stares have also commonly been associated with madness. In his
work on the relationship between trance and disease , Luc de Heusch enti-
tles the sequel to his chapter "Possession an d Shamanism" "Divine Mad-
ness and Hum an Re ason ." "Trance can be seen as the cultural aspect of
mental illness or as the 'ma dness of the gods,' as the Thonga put it," he
w rites (1971, 256 ). The title of Roger Bastide's last work, Le reve, la
transe et la Jolie (1972 ), is also significant, and although the autho r actu-
ally begins his book by claiming he will not "co nfuse dream, trance, and
madness, " one of the sections de vo ted to tr ance is nevertheless entitled
"Mystic Trance , Psychopathology and Psychiatr y." Jean Rouch's most
famous film on trance is called Les Maitres fous. Jean Gaborieau entitled
the second part of his study, "Ritual T ranee in the Central Himalayas, "
"Madness, Avatar, Meditation." And this is by no means the end of the
list.
E ver since Plato (and doubtless before him ), the question of the psychi -
atric aspects of trance has caused a great deal of ink to flow, if we ma y sa y
so. The debate seems to have made little headwa y, however, and I shall
not , of course, enter into it. First, because it is absolutely outside my com-
petence; second, because the only val id way of doing so wo uld require an
interdisciplinar y approach. All the same, it is of rwofoid interest to us.
First for its own sake, and because it is fundame nt al; second, because the
psychiatric problem posed by trance is closely linked to that of the rela -
tionship between its psychology and its ph ysiolog y. And it is at the junc -
ture of these two components of tr ance chat the problem of the "effects" of
music lies, at least in part. It must be acknowledged that, for an undertak-
ing such as mine, the absence of any general theor y of trance as a reference
point does nor simplif y the cask.
Seen from my perspective , the problem of trance is clearly a "cross-
culrural" one, as the Americans say . Bur to the extent to which they deal
w ith hysteria, the very w riting s on trance also raise in and of themsel ves a
cross -cultural problem . In the U nited Scares, for instance, the vvords "hys -
reria" 36 and, more recently, "neurosis," 37 have been erased from the psy-
ch iatric lexicon. Yer in France this is far from being the case; there hysteria
as a concept continues to have a he althy livelihood, both in ps ychoanaly-
sis and in psychiatry. 38All the same, opinions on the psych iatric nature of
trance could nor be more opposed. Subsequent to many other writers,
Georges De vereux 39 peremptoril y declares that there is every reason to
"s ee the shaman as suffering from serious neurosis, or even as a psychotic
in a state of temporar y remission." Roger Basride, on the other hand
(19 72, 67 ), supports Hersko virs's thesis "w hich considers mystic trances
as 'normal' cultural phenomena in certain socieries."' 0 I. M. Lewis, for his
II!"
pt:
16 CHAP T ER ONE

part, in his chapter entitled "Possess ion and Psychiatry" ( 1971, 186, 196 )
proposes char shamans and p ossessed pe rso ns be class ified am ong "th e
mass of ordinary 'no rm ally' neurotic peop le." Is trance a no rm al p heno m-
enon or a pathological one? Although he has not offered a solut ion to th is
ques t ion, Claude Levi -Strauss has at least prov ided a clea r formul at ion of
the problem . Scarcely abridged, rhe essential pa rt of what he w rites follows.
H av ing observed that "it is convenient to compare the sh ama n in hi s
trance state , or the protagon ist in a scene of possess ion, to a neu ro tic,"
and that "in societies where scenes of possess ion are prevalent, possess ion
is a form of behavio r open to all .. . ," he poses the ques t ion: "On w h at
grounds may we assert that ind ividuals corresponding to t he ave rage of
their group .. . and occasionally man ifesting a sig ni ficant and ap prove d
fo rm of behavior, should be tr eated as abnormal?" T o wh ich he gives th is
answer ( 1960, xv iii- xix ):

The contradict ion I have just spelled out can be resol ved in two different
ways. Either the forms of behavior described by the terms "trance"
and "possession" ha ve nothing to do with those th a t we, in our society,
call psychopathological; or one ma y regard them as being of the
same type, in which case it is the connection with pathological states
that must be regarded as contingent, and as resulting from a condition
particular to the societ y in which we live . In the latter case we would
be faced by a further choice of al tern ati ves : either so -called mental
illnesses, in realit y outside the realm of medicine, must be regarded
as sociological events affecting the behavior of indi\·iduals who have
been dissociated from the group in a particular \\·ay by their personal
histories and constitutions, or we must rec og nize in these sick people the
presence of a state that is rruh- path ological but of physiological origin,
a state that tends to create a terrain fa\·orable, or. if you \,·ish,
'sensitiz ing ,' to certain svmbolic forms of beha\·ior that still fall solely
\\·ithin the realm of sociological interpret a tion.

A relativist theor y, then, and one that moreover leaves open the poss ibil -
ity of "one da y seeing ph ysiologists disco ver a biochemical substratum for
neuroses," without it being invalidated thereby.' '
In the United Stares there are those who have attempted to resolve the
problem by resorting to a psychiatrically neutral concept, that of "altered
stares of consc iousness."" According to this line of though t , trance,
whether it is shamanic, possessional, mediumistic, o r of any oth er kind,
would be just one "altered st are of consciousness" among many ot hers,
and th is term would cover nor onl y that which ethnograp h ica l lite r atur e
identifies as "sta tes of dissocia ti on or fugue, hyste ria, halluci n atio ns, cata -
lep sy, ep ilepsy, hypnosis, somnambulism, and so on," but also suc h st ares
as those induced by highwa y hypnos is or bra inwashing on t he o ne ha nd
and possession by the Hol y Ghost on the other (Bourguigno n 1973, 4, 8 ).

L
T ranee and Possession 17
To consider dreaming as well as just anothe r state of alte red co nscio u s-
ness, as Erika Bourguignon herself does, can only inspire doubt as to the
va lidit y of this concept. M y feeling is that it actually is not a con cept at all,
but that it only serves as a convenient means of sidestepping the proble m .
If the purpose of using it is to avoid "committing" oneself to "any part icu-
lar explanatory theor y" ro account for the trance state (ib id. , 5 ) and falling
into "a mateur di ag nosis" based on the more o r less appropriate use of the
language of psychopathology (and this is indeed the dange r ), then wh y
not simply content oneself w ith calling a tr ance a trance and leave it at
th at? This is the solution I shall adopt here, alt hough I ne vertheless rese r ve
the right to return briefl y to the question of hys teria in a later ch apter.
In ethnology, one ma y say, evaluating the right o r w rong of relegat ing
trance to one or another category of contemporar y Western nosography is
nor of crucial importance. The real interest lies in finding out the categor y
to which the "native" assigns it within his own system of thought. Certa in-
ly' But it so happens, first of all, that in the great maj o rit y of the cases to
be examined here , inform ation on this point is either nil or unrel iable.
Co nsequently, any attem pt to generalize in thi s domain would be vai n.
Second, we shall see right away th at even if inform ation is ava ilab le , and
even if it is actually reliable, this does no t necessaril y help matters.
Gaborieau (1975, 165 ), in his stud y of trance in the Him alayas from
,,·hich I quoted earlier, tell s us th at the possessed per on is referred to as
the god's "beast of burden," and that ro describe the "psycho logica l trans-
formatio n " he undergoes whe n he enters into tr ance "the y say that he be -
comes mad (baulino ) or that he is mad (bau/0 )."' 3 "A person is referred to
as mad," Gaborieau obse rves, "when he can no longer carry out his famil-
ial and social obligations." H e then adds : "The same term is applied to the
ethnolog ist who, arri ving from who knows whe re, with no famil ia l or so-
cial insertion, occupies the same marginal position as the madmen belong-
ing to the soc iet y he is studying." Such inform at ion invites us to recon sider
the whole situation so r adically th at we had better forger abo ut it on the
spot!

The Representations of T r anee : Shamanism and Possession

Alth ough trance can manifest itself in a profane context - and we sh all
treat such an example among the Arabs in th e second part of this book -
throughout the wor ld it is usually assoc iated wi th reli gion. Religious
trance will therefore be our main concern from now on . This said, it oc-
cur s in such a va riet y of forms and is the o bject of such a multiplicity of
representations" th at it wo uld be impossible to establish its rel ations with
music - o r indeed w ith anyt hing at all - if one did nor begin by making a
few broad distinction s w ithin such a vas t domain.
18 CHAPTER ONE

Everyo ne wi ll agree th at tra nce is associated, in an ov erwh elm ing n um -

ber of cases, w ith two aspects of religion: sha m anism and p oss ession . But
as we sh all see late r, the relations between music and sh am anic tr ance on
the one hand, and possession trance on the other, are by no me ans the
same. It is therefore essential to establ ish clearl y, from the out set , w h at
differentiates the two - and needless to say, to do so w ithout re sort ing to
music since this would make our argument circula r.
Shamanism and possess ion have so man y po ints in common -tr ance, in
particular - that the d istinctio n between them is not always self-evident,
to say the least. Certa in authors - M ircea Eliade (1 951 ), Luc de H eusch
( 1971 )- think it is very important to distinguish between them. Oth ers on
the contrary, and most notabl y I. M. Lew is ( 1971 ), think there is no bas is
for such a distinction and tha t there is therefore no re ason to draw o ne.
Others sti ll- Er ika Bourgu ignon ( 1968 ), for example -re cog ni ze the dis -
tinction but blur it with so man y nuances that it almost disap pears . In my
op inion shamanic trance and possess ion trance constitute two very diffe r-
ent - and indeed opposite - ty p es of relationship with the inv isib le, and
thi s is why the y give rise to manifestations as different from one another
as, for example, a shaman ic sea nce in central Asia and a ta rant ism seance
in southern Ital y.
Borrowed in the seventeenth centur y from the T ungus, "a people, sm all
in numbers (25,000 ) bu t scattered across the w hole of Siberia fro m the
Jenissei to the Amur and from Lake Baikal to the Tamir pe nin sula,"" the
word "shaman " has graduall y achieved w idespread dissem ination in eth -
nological wri t ings, and its use, once restricted solely to Siberia, now ex -
tends not onl y all the way to Southeast Asia and Melanesia but also, via
the Arctic, through the two Americas as far as T ierra de! Fuego. As a re-
sult , the word "shamanism" has come to include religious pr acti ces of the
greatest diversit y, yet all presenting - at leas t in principle - common fea-
tures . Seen from the perspecti ve of this book , what are these feat ur es? Fo r
my answer tO this question , I shall turn to Mircea Eli ade and his work on
sh am anism. 06 Howe ver disputed certa in p arts of it ma y be ," th is work
ne vertheless is the only available, general synthesis of this vast do m ain.
All the same, the example s I sh all borrow from it are ent irely dissoc iable
from Eliade 's overall interpretati o n of shamanism .
In the ch apter describin g shamanism in central and northern Asia ,
which deals successively w ith the Tart ars and the Kirghiz, the Samo yeds ,
the Ostiak and the Lap p s, the Burya ts an d the Yakuts, the Chukchi, the
Yukagh ir and the T ungus , with all their va rious subgroups , besides "sha-
man" - naturall y- and along w ith "ecstas y" (or "trance "),' 8 "chant, " and
"drum ," the tw o wo rds th at recur most frequentl y are "jo urne y" and "spir -
it. " T he shaman's trance is thus conceived as a journe y undertaken in the
comp any of the spirits he embodie s. The sha m an's soul leaves his bod y
Trance and Possession 19
and voyages through the invisible regions in order to meet the dead or the
spirits, and his journey usuall y takes the form either of an "ascent to heav-
en" or of a "descent into hell, " to use Eliade 's wo rds , or better yet, of an in -
cursion into the "upper " or "lowe r" worlds. "Among the Xingan Tungus,
the shaman vigoro usly strikes hi s drum three times and raises it in order to
indicate he is ascending into the upper world. He strikes it three times and
lowers it when he is descending into the lower world. "• 9
Among the Eskimos , it is in the "depths of the ocean" that the shaman
undertakes his journey "to seek our his patient's soul and bring it back to
his or her bod y." In addition, he is a "sp ecialist in magical flight. Some sha-
mans ha ve visited the moon, others have circled the earth," Eliade writes
(1968 , 223 ). Among the American Indians , shamanism also rakes the
form of a journey. Besides the examples mentioned by Eliade , an account
of such a shamanic journey, collected among the Cuna Indians of Panama,
made famous by Levi -Str auss's interpretation (1949 ) should be cited here .
The same situation exists in Southeast Asia and the Indian archipelago. 50
Nothing of this kind occurs in what is by general agreeme nt termed pos -
session. In possession the communica tion bet ween visib le and invisible
worlds operates in quite the opposite direction: it is no longer the man
who visits the inhabitants of the invisible wo rld bur, on the contrary, the
latter who visit him . One could perhaps say, by oversimplifying the oppo -
sition, that since rel igious trances are essentiall y private and dramatic
communications with the dead, the spirit s, or the gods -depending on the
particular reli gious system concerned -men disp ose of two major means
for achieving such communications: either they ha ve to go to the spirits,
or the spirits ha ve to come to them. Whether we examine menadism in an-
cient Greece or demoniac possession during the Renaissance, the ziir cult
of Ethiopia o r the rab cult of Senegal, the orisha and vodun cults in the
Gulf of Guinea and Brazil, basangu in Zambia or hdu bong in Vietnam, or
any of the various forms of trance to be found in Bali , nowhere do we find
an y evidence of trance being viewed as a journe y made by man into the
spirit world. In every case it is int erpr eted as invol ving the arrival of a spirit
or a god in the world of men . This difference is a radical one: in the first
case we ha ve shaman ism ; in the second, po ssess ion . Since in both cases
the event is experienced from the viewpoint of the human involved, in the
first case a journe y is taken and in the second a visit is recei ved -consider-
able distinction indeed. 5 '
Starting from these two basic schemes, howe ver, man y combinations
can be constructed. There is nothing to prevent someone, for example,
from setting our on his o r her journey only after ha ving "incorporated"
auxilia ry spirits to ass ist him or her in the other wo rld . 52 This is in fact
w h at happens in shamanism, so that the shaman has sometimes been de -
scribed as "possessed." In this case journey an d "possess ion " are com -
20 CHAPTERO E

bined. Arguin g from the fact th at sham anism and possession by a spirit
regularly occur together, Lew is ( 19 7 1, 51 ) is of rhe opinion that, contrary
to the views of Eliade, ' 3 and de Heusch , ' ' the di stinction between shaman -
ism and po ssessi on is "in fact untenable ." Bur this ign ores the fact that the
"journe y" component is regularl y present in trance manifestations ob-
served amo ng the Asian populations listed earlier and regularly absent
from those in my second list. This presence, or absence, is an essential fea-
ture, since it is the factor that no t only underlies all representations related
to trance bur also go verns most of its ritual and external manifestations.
Flanked by his aux iliary sp irit s, the shaman-or rather his soul - leaves
his bod y in o rd er ro trave l to the invisible world and meet its inh abitants.
During hi s tr an ce, he rec ounts the events occ urring during his journe y,
which is often undertaken with the aim of recuper atin g a sick person 's
soul. "The m ain function of the shaman in central and no rth ern Asia is
m ag ical he aling, " Eli ade obser ves (1968, 179 ). Hi s "jo urne y" is for him
the supreme method of exercizing th at function. Hi s "possession" by hi s
gu ardi an or auxiliar y spirit s is ultim ately no m ore than a seconda ry aspect
of it.
One ma y object th at there a re cases, notabl y among the Tungu s, in
which th e shaman "incorpora tes his familia r spi rit s and undergoes a sha-
manic trance wi th ou t going on a jou rn ey" (E liade 1968, 199 ). In repl y, it
mu st be noted that in the functionin g of any institution, the logic of things
accentu ates some tim es thi s, sometime s that featu re. " The fact still re-
mains that amo ng such peop les the "descent inro hell"- wh ich in th eir
"pusillanimit y" present-d ay shamans "do nor d are to undertake" (Eliade
1968 , 195 )-i s the great mode l of the shaman ic adventure. ' 6 Laurence De-
la by, for her p art , also stresses "the import ance att ri buted by the Tungus
ro the perilous jou rne y int o the lower world th at only the gre at shamans
w ere abl e to m ak e" (1976, 13 0 ).
Let us n ow mo ve on to rhe role pl ayed by th e in co rp ora tion of spirits in
shaman ic tr anc e, and consider whethe r thi s inco rpor at ion occurs in the
same way in shamanism and in possession . s; We shall find that it emphati -
cally does not, and that once aga in the rel ations inv olv ed in the two phe-
nomen a are diametricall y oppos ed. H ow is such incorporation seen in
shamanism? A br ief summ a ry of what happens amon g the Gold ' 8 will
provide ve ry significant insights. As amo ng other Siberian tribe s, the Gold
shaman is a "psychopomp"; in other wo rds , he is the one who guid es the
souls of th e de ad int o the lowe r world. Each person who d ies has two fu-
ner al ceremonies. Th e second , wh ich is the more imp ortant, "takes pl ace
so me time after th e first " and "terminates w ith the soul being led to the un-
derworld ." Based on Eliade's description (1968, 17 4- 7 6), I shall no w pro-
vide a summa ry of this cerem ony .
Th e funera l ceremony lasts seve ral d ays . T he first day, a fter donning hi s
cos tume , th e shaman rake s up his drum
T ranee and Possession 21
and sets our in search of the soul in the area around the yun [the rem of
central Asian nomads]. While searching, he dances and recounts rhe
difficulties that lie ahead on rhe p ath to the Unde rwo rld. Finally he
captures the soul and brings ir back into the dwelling ... . The feast
continues late into rhe night ... . Th e next day, he dons his costume
again and awakens the de ad person w ith his drum . .. ;[there is then ]
another feast. . . . Finall y, one morning the shaman begins his chancing
and, add re ssing rhe dead person, advises him to ear well burro drink
ve ry little, since rhe journey to the underworld is extremel y hazardous
for someone who is drunk. Ar su nset , prepar ation s are made for the
departure. The shaman chants, dances, and sme ars his face wi th soot.
He invokes his auxiliary spirits and asks them to gu:de rhe deceased and
himself into rhe beyond. He leaves rhe yurt fo r a few moments and
climbs up a notched rree ... ; from there he can see the path down to rhe
Und er world .. . . Returning to the yu rt, he summons rhe aid of rwo
powerful guardian spi rit s: buchu, a kind of one -footed monster with a
human bur feathered face, and koori, a longnecked bird .. .. Wirhour
the help of these rwo spirits rhe sham an could never make his way back
up from the Un der world, because rhe most difficult part of rhe return
journey can only be made sirring on rhe koori's back. After dancing and
chanting himself inro a stare of exhaustion .... the shaman calls upon
the spirits to harness rhe dogs to the sledge, then asks for a "manse rva nr "
to keep him company during the journey. Several seconds later he
"leaves" for the counrry of the dead. The ch ants he intones, rhe words he
exchanges with rhe "manservant" enable those present to follow his
itiner a ry .... The y arrive in rhe Underworld . The dead at once cluster
aroun d and demand to know rhe name of rhe sha m an and of the
newcomer. The shaman is careful not to utter his real name . He seeks
among the crowd for the spirits of rhe kinsmen who were closest during
their lives to the soul he has brought with him and consigns it to their
care. He then return s to the land of the living as quickly as possible and,
once back, recounts ar length all rhat he has seen in the land of rhe dead .

I oug ht t0 add, since Eliade takes it so much for granted he does not men -
tion it , that throughout this time the shaman is in a trance. 59
othing could be more different from this account than trance as it is
obse rved in Africa. Among the Gold the shaman acts as the master of the
spirits he embodies: he uses them as "gu ides, " he sits on the back of one of
them, he makes them harness his dogs, he converses w ith his "m anser -
vant." In o rder t0 return from the underworld, as we saw, he rides one of
his auxiliar y spirits as though it were a ho rse. Among the Ethiopi ans of
Gondar, on the cont rary, as well as among the Songh ay of 1iame y, the
possessed person - or "possessee" - is the one who is ridden by the genius!
As Eliade says (196 8, 89 ), when sh am ans among the Chukchi, the Eski -
mos, the Lapp s, and the Seman g tran sfo rm into a wolf, be ar, reindee r ,
fish, or tiger and imitate these animals ' action s and voices, they are not be -
22 CHAPTER ONE

ing "possessed by these creatures; rather, the shaman takes possession of

his auxiliary spirits: he is the one who transforms himself6° into an an im al , ·
in the same way that he achieves a similar result by putti ng o n an ani ma l
mask .... " In Afr ica, the possessee does not imitate the god he embod ies,
he is that god. In Asia, the shaman coexists with the gua rdian or aux ilia ry
spirits he has embodied by means of his magic chant, and this is w hy it is
possible for him to converse with them. Unless I am mistaken, in Africa
(at least in all of the cults I ha ve mentioned ) such coex istence, w ith in one
and the same individual, of the subject in trance and the spi rit possess ing
him, never occurs. The persona of the divinit y is substituted for that of the
subject but does not coexist with it. Any dialogue between the two would
be unthinkable.
Possession, or more precisely the possession fit as obse rved so fre quent-
ly in Africa (a subject we shall subsequently discuss at length ) can neve r-
theless be found among Siberian peoples who practice shaman ism, and
particularly among the T ungus. In these cases we are dealing with dist ur -
bances brought about in certain individuals by spirits who tor m ent t he m .
These possessed individuals, who are "for the most part frustrated or dis -
appointed women" on whom t he "visitation of spirits" confers "a certa in
social importance" (Delaby 1976, 28 ), are not called shamans but some -
thing quite different. However, as with the possessed person in Africa, it is
frequently through disturbances of th is order that the shamanistic voca-
tion first manifests itself, or, more correctly, as Delab y puts it (1976, 37 ),
that the future shaman is elected. Shamans, "when they begin," thus also
fall into the category of the possessee , but "they transcend that category
by master ing the spirits that cormenc chem" (ibid., 28 ). Once embodied
within them, the spirit of che shamanic ancestor who has chosen them
(ibid., 38 ) becomes their principal auxiliar y spirit. From chen on, che sha -
man is the one who "summons" chat spirit, together with all those fellow
spirits that ordinaril y accompan y him, and who embodies him "at will"
and expels him "when he has no further need of i1is ser vices" (ibid., 28 ).6 1
Master over che spirits he embodies, che shaman is also in control of his
trance , another factor that helps us co distinguish him from the possessee
who, as we shall see lacer, merel y undergoes this state. Thus it has recentl y
been proposed chat we define che shaman as someone "who can enter at
w ill into a non -ordinar y ps ychic state (in which either his soul undertakes
a journey inco the spirit world or he himself becomes possessed by a spirit )
for che purpose of making contact with che spirit world on behalf of mem-
bers of his community. " 6 2
There are cases, however (and the Kahm-Magar in the H imalayas,
filmed and described by Michael Oppitz [1981 ) are an example ), in which
in the course of the same ritual the entranced person appears sometimes as
a shaman who vo yages and controls his trance , sometimes as a possessed
Trance and Possession 23
person who totally submits to the will of the possessing spirit, identifying
with him but not mastering him at all, quite the contrar y. Shamanic trance
and possession tr ance - this time in the full sen se of the wo rd-can thus
alternate in one and the same person. Or, if one prefers, one and the same
person can undergo in succession these two forms of trance . It seems that
no one can experience them simultaneousl y. The distinction to be made •''I
between shamanism an d pos session, consequentl y, remains vali d. The dif -
ference between the shamanic and possessi on tr ance thus seems to rest on
three factors: the former is a journey m ade by a man to visit the spirits, the
latter is a vis it by a spirit (or divinity ) to the world of men; in the former
the tr ance subject gains control over the spirit embodied wi thin him , in the
latter the reverse is true; and lastly, the former is a volu nt ary trance w here-
as the latter is an in volunta ry one. 63
Howe ver, here, as in so man y other fields, displacements can occur
from one model to the other, hybrid forms arise, contacts are made or
lost, giving birth to mixtures that are continually done and undone. This
is why the configurations of religious trance are so extremely va ried , and
also why, in man y cases, it is difficult to be sure whether one is dealing
with shamanism or with possession. In this respect, based on an anal ysis
slightly different from mine, and one to which I do not wholly subscribe,
Luc de Heusch 6 " described the complexity of thi s situation very clea rly. It
is unnecessar y to belabor his points here. But even though there are all
sorts of possible intermediary forms bet ween shamanism and possession,
we must not forget the difference th at separates the two. It is fundamental.
In m y view , I. M. Lew is commits a great erro r in refusing to recognize !
j,
this difference. Let me give a concrete example. In his book 65 he repro-
duces a photogr aph of a priest of Ogun, the Yoruba's iron god, taken in
II
,.
Dahomey (n ow the People's Republic of Benin ) around 1950-60. I know
the priest in question , having seen him perform his functions on seve ral
occasions in 1952 when I was in the compan y of Pierre Verger , who also
pub lished several photographs of the same individual to illustrate his de-
scription of the weekly ceremon y in which this particul ar priest was the
principal officiant at the time. 66 T he caption to Lew is's photograph reads:
"The god Ogun speaks through the mouth of the shama n who incarnates
him, in Dahome y." Wh y baptize (if one ma y so express it) this priest a
"shaman"? Among the Yoruba, the trance state associated w ith th e wor -
ship of Ogun, the god of iron, has ne ver been regarded as a journey. As de -
picted , the priest in hi s trance state is absolutely not recounti ng what is
happening to him in the otherworld. Nor is he imit atin g one of the gua rd-
an or auxiliar y spirits whom he has embodied as the resu lt of his sum-
nons. O gu n is w ithin him and he identifies himself to Ogun , whose sym-
Jols - iron cutlass and bell-he wears, but he is so little master of the god
hat he is accompanied by a priestess whose function is "to cal m Ogun
 24 CHAPTER 0. E

when he becomes too violent (Verger 1957 , 151 ). "Ogun kp<Ilft o"-
"Ogun gently!" the choir of women sing when the divinity takes posses-
sion of his follower too abruptly (ibid., 195 ). When the shaman in his
trance returns from his journe y to the under world, bringing news of the
dead he has just encountered there, he, the shaman, is the one who speaks.
When the priest of Ogun, as seen in thi s particu lar photograph, add resses
himself to the people around him, it is not he, th e priest , who speaks; it is
Ogun. To say in this instance that Ogun "is speaking through the mouth "
of the individual embodying him, as Lewis does, is to suggest th at Ogun
and the priest are coexisting inside one and the same person, as does in
fact happen wit h the shaman . But thi s is absolutely nor th e case. The per-
sonal it y of the priest as priest has been totall y replaced by that of Ogun.
The chorus, which continues its chanting throu ghout the priest 's dance,
sings about Ogun and Ogun alone . To view this priest as a shaman is com-
pletely to misunderstand the representations th at underlie his trance and
to create a wholly unnecessar y confusion in a situ ation that is al re ady
quite complicated.
The responsibilit y for this confusion does nor fall entirel y on I. M. Lew -
is, who is onl y perpetuating that m ade long before by S. F. Nadel in his
well-known article, "A Stud y of Sh amanism in the Nuba Mountains"
( 194 6). adel presents the cult he obser ved among the Nuba of Kordofan
as corresponding "in all essentials to the classic al shamanism of Central
Asia and North West Africa " because , like the latter, '·it rests on the belief
that spirits may possess~uman beings and on the practice of establishing
communication w ith the supernatural th rough human be ings so pos -
sessed" (19 46 , 25 ). In fact, it constitutes a case of possession at the service
of mediumship , and conforms in everv res pect w it h the case described a
few pages further on in this book.
Thi s is not to say that tr ance in bl ack Africa onlv has one pos sible cause,
which is spirit possession. In Afric a there a re in fact tr ance s not attributed
to a spirit. Such is the case, for instance, among the Zande , in Central Af-
rica. The witch doctor, '\\"110 is both di viner and mag ician" (Evans-
Prirchard 19 3 7, 14 9) and whose task "in both roles is to cou nteract \\·itch -
craft," perform s these acti vities in "seances '· durin g wh ich he is entranced
("H ow Wirch-doctors Conduct a Seance,·, ibi d., 14 7- 82). Eva n s-
Pritch ard 's description leaves no doubt on this point, even though he him-
self onl y dare s say, prudently, that th e w itch doctor ..spe aks as th oug h in a
trance" (ibid., 165 ) or "is now entering in a tr ance -like co ndition " (ibid.,
167 ). To what do the Zande attr ibut e this st ate) Eva ns-Pritchard does nor
pro vide the answer. H ad we nor knO\\·n it \\·as in Africa, howe ver , we
would ha ve belie ved th at the seance was raking pl ace in Siberia and that
the witch doctor was indeed a sh aman. But of cour e to be val id such a n
interpre tat ion wo uld require much more information.

. . . ·._·· f':ri~::~::}i:ijf{tt;{·I~-:{:;"·.::}~
:·~"_:,·::'-;f(·:"~:}t~~}~~4j
. - -. ~ - . -_. ; ·~·.- 2,~·;.-._-.
-\. -~ :.:~-;!~~;t
~~4~-;:~:;~)~3~:f~:-~~i:}~~~~~~:)-~~~J,f~·:·45~::?J·~~\~
 T ranee and Possession 25
Other fo rm s of trance in Africa r aise as well problem s of typolog y. I am
thinking here of those assoc iated w it h va riou s cults in equatorial Africa
(the Mbwiti in Gabon, n otably ), in which taking hallucinatory drugs pl ays
an import ant role. But I shall limit myself to mentioning th e fact, for fea r
of bein g led too far away.
Let us now return to the essential point. W hat we need to remembe r
abo ve all is th at shaman ism and possession are the products of two quite
d iffer ent ide ologies of trance. As we shall see later on, these two ideologies
co rresp ond - most logicall y- to two quite d ifferent uses of music.

Possession , Inspiration, Communion : Idenrificator y Tr ance

T o speak of personalit y subs tituti on, as I have just done in the case of th e
priest of Ogun, to see it as a relationship "between a spi rit thought of as
ac tin g and an indi vidu al thoug ht of as acted upon," as Michel Leiris
(19 58, 78 ) puts it with reference to th e ziir cult, and to consider th at this
constitutes a relationship of "possession," certainly is, as Leiri s says, a "re-
duct ive qualification." Nothing could be more haz y, less sha rpl v foc used
than the concept of possess ion, so that desp ite its great convenience it is
also the source of a gr/at man y confusions. H ere aga in, several broad dis-
t inc ti ons need to be ma de.
Jeanne des Anges, the most notorious of the possessed nuns of Loudun,
was described ·b y everyone, including herself, as "possessed ." Jean Cava -
lier , the famou.s Cam isard, was described by himself and everyone else as
"ins pir ed" - w1i:h the exception of the papists, who described him as pos -
sessed in order to make people believe that the devil was in fact possessing
him. We can apply neither of these terms to the Sufi in tr ance : "possessed"
wo uld me an that he was so by either God or th e dev il, which wou ld be er -
roneous; "inspired " would impl y th at his sta te was due to the presence of
the Hol y Ghos t , wh ich is a religious concept fo reign to Islam. He is in the
state of ha ving found (wajd )6 ' God. Leiris (1958, 78 ) refers to adepts of
the ziir cult in Ethiopi a as being p ossessed bec ause the Amharic vocabu-
lar y used there does in fact convey "t he ver y general ide a of seizure of the
spi rit by force - the exact equivalent of our 'possession ."' When an adept
of the vodun cult in Benin enters int o a trance, t he vodun is said to have
"co me in to his head." 6 8 In Brazil, the candomble adept is sai d to be "in a
state of santo" (Binon 1970, 163 ). In the case of the H aus a bori, in Niger,
the woman in tr ance is sa id to be the "mare " of the god, o r to "hav e the
gods on her head," o r again, to be "in bori," \,·hich is to say in a state of
ebullition (M onfouga -Nicholas 1972, 193 ). T he list could be much long -
er stil l, but let us stop there. T he question is to understand how the rela-
ti o nsh ip between t he subject in trance and the divinity resp ons ible for that
st ate is conceived .

-~!,,~- ~ ·,--,:'~,!<'..~~;;l?_
~; ~§.'l;'.~~~~~ _.,_""~:~{~~~~-,-...~~- ..,..}
. .

•
•
~ '
·.
-- · ·--
J
;,,i-';, -
~.
--~·,··~
-v•"
_,.-
•>;
,~-
./.-,,
.
:,; ·,,i'h'IOS"; '"")-..,;;,
-- ... -·· ~ .,."l-,:
.··:;'.il~.,/)~
,;;,~r-',;: _.,,?c-.'..'.;-,:<.s~;j'/;it,
< •
?/"_
<::<<_:'}-
-
'~:~~-;:;;::Ji·i-t~,-·,·f:(
jl,"f1+;;,~~-,'l:!,'·.:-•··
,•t,;."'-''Ci:·jf§f/{.l
""'-~~.....r~. ·~iiij;it~,.U-
26 CHAPTER ONE

In the case of a relationsh ip of t his kind, ho wev er, we must not expect
to find very cle arcut notions capable of pro viding us with precise and
clearly defined categories. By definition, we are confronted here w ith the
domain of the un speakable , the vague, the subjecti ve, and the emotional.
Nevertheless, by taking as our base the vocabu lary used ro designate this
re la tion sh ip , the behavior by wh ich it is expre ssed, an d the internal logic
of the re ligion of wh ich it fo rm s one as pect , we are able ro distinguish
three broad types o f mystic tr anc:e .
T he first is characterized by the fac t th at , during tr ance, the subject is
thought ro have acquire d a different personalit y : th at of a god, spirit, ge-
nius, or ancesror - for wh ich we m ay use the gen eral term "deit y"- who
has taken possession of the su bject, subs tituted itself for him , an d is n ow
ac tin g in that subject 's place . In a later chapter we shall return ro the m an-
ner in which this represe nt ation o f things is often founded on the ide a th at
an individual possesses seve ral souls, so th at , using the bori of the Niger
H ausa as an example, ther e is one soul th at "the gods send away durin g
possession in order ro rake its place" (Monfouga - icolas 1972, 194 ). For
a longer or shorte r period the subject then becomes t he god. He is the god .
We ca n call this possession in th e stri ct sense of the word.
In the second categor y, rather than having switched persona litie s, the
subject is thought ro ha ve been invested by the deit y, or by a force emanat-
ing from it, which then coexists in so me wa y with the sub ject but never -
theless controls him and causes him to act and speak in its na me . The mo st
frequent ex a mple of this relationship is that found in trances at tributed ro
the Hol y Ghost. I sh all refer ro this categ o ry of tr ance not as p ossessio n ,
but as inspiration.
In th e third categor y the rel ationship between divin ity an d sub ject is
seen as an encounter which , dep ending upon the individual, is experi -
enced as a commun ion , a revelation , o r an illumination. Unlike the pre-
vious two relationships, this one does not invol ve embod iment of any
kind. I shall therefore refer tO the trance state brought about by w ajd, and
achie ved by the pr actice of dhikr, which will be described in the seco nd
part of this book, as communion or "com muni al" tr ance. For a Sufi , or for
an y Muslim, there can obviousl y be no questi o n of embody ing Allah, o r
of identi fying oneself with him.
H avi ng defined the se three categories - not as abstract models but on
the basis of concrete examples - let me hasten ro add that, in re ality, thing s
are not always as simple as this division might suggest, and th at it is often
difficult ro kn ow wit h w hich of the three categories one is dealing . To re -
turn ro the case of the Sufi , it can h appen , con tr ary to all the principles of
the Muslim faith, th at it does take the form of an indisputable embodi -
ment. Amon g other example s, there is this description of a dhikr sean ce by
a Fez derq;;va vota ry 6 9 in wh ich , speak ing of God (Allah ), the tr ance sub-
T ranee-and Possession 27
ject cr ies out: "H e is in me! H e is in me! H uwa! H uwa! [mea nin g He! He! ]"
Man y people wo uld say that this, whether on e wishes to acce pt it o r no t , is
purely and simply a case of possession . For m y part, I shall say that it is in-
spira t ion, but that in any event more than me re communion is clearly in-
volved, contra ry to what the logic of Sufism m ight lead us to ex pect. In
other cases the distinction I ha ve suggested between "possession " and "in-
spiration" - which implies a differentiation between two different modali -
ties of embodiment - will not seem either self -evident or necessary to ev-
eryone. T he Shakers of Saint Vincent in the Antilles reg ard their tr ances as
resulting from the presence of the Hol y Ghost . The y say the y are "shaken"
by him (Henne y 1973 , 238 ), but the y do not ident ify t hemselves w ith him
in any way. Jean nette H enney ne vertheless explains these t rances as res ult -
ing from possession by the Hol y Ghost. Fu rt her justification of my point
of view is therefore necessary , and this brings us back to the identificatory
aspec t of possession .
Let us take the case of the candomble , pr act iced in Bahia, Br azil. T he
divinities (orlxa ) ve nerated there form p art of an extremely populated
pantheon; Roger Bastide (1958, 141 ) names ten wi thout claiming his list
is complete, a nd each of th ose on the list is plural. E ach of these de ities (a ll
of African origin ), Ogun (god of iron ), Yemanja (goddess of fresh and sal t
water ), Sh ango (god of lightning ), an d so on, is capable of incarnat ing
him or herself into an adept, and thereb y bringing on the trance st ate . Th e
ade pt then identifies himself wi th the divinit y and manifests this ide nti fi-
cation by carrying the deity's symbols (a double -headed ax for Xango, a
sword for Ogun, an articulated fish for Yemanja , and so on ), by assuming
his or her humor, by imitating his or her beha vior , and by executi ng dance
steps expressing the deity's indi vidu al characteristics. 70 One must adm it
that in a cu lt in wh ich possession can be attributed to a va rie ty of deities , it
is essential to kno w w hich one is involved, hence the nece ssity to ide nt ify
him o r her . The same can be said of many other cults. To name but a few,
there is the Vietnamese hdit bong, the Ethiopi an zar cult, and that of rab
in Senegal. Ident ifying the sp irit o r divinity resp o nsible for the po ssession
occup ies a very imp o rt ant pl ace in the initiation rit es of the se cults: unless
one knows to w hich deity any particular possession should be attributed,
hO\\. is one to address him, m ake rhe cor rect sac ri fices , and gener ally be -
have toward him in the appropr iate manner' In the case of tarantism -
which is nor, properly speak ing, a cult -i d entif ying the particular spider
whose bite is resp onsible fo r the trance sta re constitutes a crucial m o ment
of the ritu al, since here aga in there are sever al different spider s th at ma y
have bitten rhe sick person. The first thing to do , therefore, is to ascertain
whic h one is in vo lved (a process carried o ut w ith the aid of music ). More -
ove r , one of the dance figures through w h ich the tr ance is expressed con -
sist s in imit ating the spider by walk ing, if th at is the right wo rd, on all
 28 CHAPTER ONE

fours with one's back to the ground. So th ere are trances w hose ritual is
identifi cato r y in nature, an d in a double sen se, moreover: on the one hand
it is essential, before anything else, to identif y the deit y responsible for the
perturbation, and on the other, once this has been done, it is the n neces-
sary ro become identified with him or her. I shall say, therefore, that there
is possession when the ritu al is identificarory, an d only in this case.
Nothing of this kin d occurs in what I have proposed we call inspi r ation
or communion trances, even though in one of them there may be embod i-
ment (as in possession ) and not in the other. When the dervish, as a re sul t
of dhikr or samZi', " falls into a trance and dances, he is not identifying
himself in any way with Allah, even when - an extreme case inadmissible
to the purists - he cries our "He is in me' " Neither (on the orthodox side )
the followers of the Shlustes (a Russian flagellant sect ), 72 who dance in a
trance induced by the presence of the Hol y Ghost, nor (on the Protestant
side ) the Camisards of rhe Ceve n nes (Fr ance ), who are moved to prophes y
by inspiration, identif y themselves w ith God o r the H oly Ghost. The lat-
ter causes them ro shake and dance, o r talks through their mouths; bur
those he insp ir es do nor in any way imitate him. Would imitat ing Allah,
Jehovah, o r the Hol y Ghos t be conceivable' The same may be said of Eli-
sha 's trance in the Bible , when he feels the hand of Jeho vah upon him be-
fore he begins to prophesy (2 Kings 3: 10- 15 ): he senses rhe presence of
God upon or within him, but there is no question of identifying with Him
o r imit at ing Him.
Thu s nonidentificatory tr ance inspir ation or communion trance ap-
pears to be characrerisric of Islam, Christianity, and Judaism. In other
words, it seems to be linked wit h the log ic of religions of transcendence . It
nonetheless signals the presence of a divinity. It is an epiphany, and in thi s
respect it should be ranked alongside possession, even though the latter is
identificaror y, rather than alongside shaman ism, even though shamanism
is also nonidentificarory. We shall see later char identificatory and non-
identificarory trances each make very different use of music. This is why I
have insisted on the importance of making thi s distinction.

Di vinaro-ry T ranee and Mediumship

Shamani m, possession, and inspiration have in common the fact that
rhe\· all - each in its own w ay - usually lead to di vination . Di vinator y
tr ance is nor their prime goal, but it does constitute a natural extension of
such states. Shamans are not specialists in divination, any more than those
who are possessed or inspired, but while thev are in a trance they are
asked questions, their ad\· ice is solicited, and the y are requested to predict
the future. Besides this kind of trance in which divination occas ionally fig-
ures, there is also another type in wh ich divination is the sole purpose. I
shall label this type mediumistic trance or mediumship.

L
T ranee and Possession 29
Mediumistic trance also manifests itself in numerous ways. In Greek
antiquity, the most famous example was the Pythia of Delphi. As we
know, she was inspired by Apollo, but her trance did not in any way in-
vol ve identification with or imitation of the god. People came to consult
her and she simpl y delivered her oracles. In Tibet the so -calle d govern-
ment oracle has a comp ar able function, but in this case the trance is trig-
gered, a nd manifests itself, in a very different way; I shall have occasion to
refer to it several times later. There a lso are numerous examples of medi-
umistic trance in both Africa and Asia. Since by its ver y function it is trig-
ge red on command (frequentl y at close int erval s), it is only logical that it
wou ld often m anifest itself in fairly nonviolent ways, and that on occa-
sio n, it wo uld occur so discreetly as robe imperc ept ibl e. This is the case
amo ng the Mofu of C ameroon, for example , where one finds both male
and female soothsayers. J.-Fr. Vincent ( 1971, 108 ) writes that the women
"identif y w ith their genius tO the point of sa ying 'I' or 'we ' instead of using
th e third per son." Their tr ance m anifests itself in an extremel y reserved
manne r. The men, who practice divination in what one might call a more
profess ional way, are described as being "inspired " (ibid. , 7 6 , 77, 104 ),
and th eir possession by the "gen ii of devination" is thought to be a "perm a-
nent state." When the y converse w ith th ese genii, an d que stion them , their
trance, if it is a trance , does not m anifest any extern al signs, and there is
nothing ro indic ate the y have entered an "al tered state" (ibid., 104 ). On
the surface, then , th ere is no trance sta te at all in these cases , even though
the soo th saye r's vocat io n was or igin ally determined by trances or posses-
sions displaying all the usual characteristics of those sta tes (ibid. ).
For all so rt s of reasons, mediumistic tr ance- wh ich one can observe
without going ro the other end of the earth - can take an almost infinite
variety of forms. It s relations wi th music do not th erefore ha ve the same
regula riti es that can be found in the case of shaman ism o r possession. I
shall devote a section of a la ter chapter to thi s subj ect.
To conclude thi s attempt at classifying religious tr ance, I should note
that we shall be dealing, a little late r on, w ith yet another type of trance ,
one closely connected wi th possession bur quire dis tinct from possession
trance, and which I feel is be st label ed, fo r reasons I shall provide later on,
as "initiato ry trance." Lastly, it hardl y need be sai d th at I do not claim ro
have exhausted all the possible aspects of tr ance in this typolog y, and that
there are many more th an those I have described. I h ave limited myself
here to those with which we w ill be co ncerned in this book.

T HE DYNAMICS OF POSSESSION
Possess ion Cults and Reprobate Poss ession
Of the va rious types of trance we h ave just investigated, possession tr ance
is the o ne th at will concern us most. It is therefore import ant for us ro in-
 30 CHAPTER ONE

vesrig are in greater derail the characteristics of possession as a religious

phenomenon. In his "Prolegomena to the Study of Possession Cults," Rog-
er Basride ( 1972, 84 ) defines such cults as sets of "functional va riab les" in
wh ich manifest and latent functions are articul ated in different ways and
give rise to a very comp lex series of combinations. Without entering into
the details of these va ri ables , let us simp ly say here that ( 1) possession
cults are forms of religion ch aracterized by a certain type of rel at ion be-
tw een a given deit y and certain categor ies of his or her zealots whom I
sha ll call adepts, chis rel ation causing the adept to become posses sed by
the di vinit y; (2 ) p oss ession itself is the socialized beha vior of an indi vidual
consisting, given certain circumstance s, in a change caking place w ithin
him , w ith the effect th at hi s usual personalit y (w hich governs his ever yda y
behavior ) is repl aced by the persona of the deit y, who dictates different
forms of beha vior to him; thi s sub stitution being accompan ied by an alter-
ation in p sychic activity generall y termed trance; (3 ) the identification
brought about in this way constitutes an alliance (sometimes reduced, one
might say, to a pact of peaceful coexistence ), the major function of which
is to lead the divinit y either to exert his po wer in favor of the possessed
person and / or his group (by incre as ing their strength, say, or protecting
them from adversi ty, or curing them of an illne ss, or revealing the future to
them ) or, alternat ively, to cease exerting his power against them .
There are also cases in w hich, far from being cultivated by society, pos-
session is reprobated either by the whole of societ y including the possessed
indi viduals themsel ves, or by one section of the population, sometimes a
majorit y, sometimes a minorit y, acco rding to circumstances. The notori-
ous possessions of Loudun , w hich were the t alk of France in the seven-
teenth centur y, were unanimousl y disappro ved of, since the y were demo 7
niacin n ature , and regarded as such not onl y by public op inion but also by
tho se subject to them . The y were nevertheless "po ssession " in the sense of
the wo rd we are using here , since the y we re identificator y : "Th ree types of
references - to the social bod y, to the angelic Orders (whence the demons
formerly fell ) and to the phys ical bod y- constitute the coordinates by
means of w hich an identity is conferred up on the se wo men," observes M.
de Certeau (1970 , 135 ). "The evil spirits wer e commanded in Lati n to
spe ak their n am es .. ." he write s elsewhe re (ibid., 28 ). We cannot help
but com p are this identific ation of the spirit by its name, in Loudun, w ith
w hat Andras Z empl eni describes as hap;.>ening among the Wolof of D a-
kar, in his ch apter on "t he naming of the rab" (the rab be ing the sp irit re-
sponsible for the possession ), nam ing wi thout w hich there could clearl y
be no p oss ibl e "ident ificato ry beh av ior" o n the part of the adept s (Zem-
pl eni 1966 , 356 ff. , 402 ff. ).
In the opinion of one of th e mo st import ant protagonists of thi s much-
publicized affai r, Father Surin (the princip al exorcist of these extraordi-

l
T ranee and Possession 31
nary vict ims of possession, before succumbing co possession himself ), the
Loudun possessions cook a truly "theatrical" form. 73 Th is does not mean
th at the y became the object of a cult. Quite the contrary in fact. Far from
cultivating possession, the point was co exorcize or put an end co it . This
explains why chis "cheater" was totally de vo id of m usic, and why we sh all
consequently not deal any further w ith this type of possession in this book
except, very briefly, in a lacer chapter devoted co exorcism.

Music and Pos sess ion: An Approach co the Problem

W id ely distributed throughout the world, possession in its cultivated
forms is associated everywhere with music and dance. Associated, more -
over, not merel y in a general way, as is the case to vary ing degrees with so
many forms of religion, but in a very particula r one, since trance, whic h is
the central phenomenon here, is often regarded as the direct result of mu-
sic an d dancing . We are therefore faced with the problem of unde rstand-
ing the origin of this power, so frequentl y accorded to mu sic, of triggering
possession trance. Of what order is it? Psychic) Somatic? Sym bolic? Effec -
tive? Imaginar y? Real? On the other hand, we shall see lacer that music,
even with in one and the same cult, may sometimes appear to be indispe n-
sible to triggering the trance and then, at ochers, quite unnecessary. Why?
How are we to explain the fact that its effect s are, as we shal l discover, so
va riable? Or that the y rake such different, sometimes even contradictory
forms from one cult to ano ther? Or chat the y have been so va riou sly inter-
preted by those who ha ve described them? There is some thin g pa radox -
ical in the relations between music and possession. Seen from afa r, in their
totality, the y appea r most of the time to be absolutely necessar y. Yet when
one looks closer, th e natu re of chat ne cessit y appears incoherent and es-
capes all attemp ts at formulating it. Thi s is because possession cults, of
which there are so many va rious types, muse be regarded ·;;isso many dis-
crete syste ms whose internal logic w ill govern the articulation of their ele-
ments differently in each case. Music occurs as one of the component
parts of chose sys tems , and the role it plays va ries wi th the models upon
w hich chose systems are structured. But , since music is itself a system, its
relations with each model are equall y determined by its own organization,
thus accounting for their protean character. Ho w can we tr y to grasp such
a constantly shifting realit y?
· Since the essential feature of possession cults is the change of state chat
occurs w ithin the adept's psyche , these cults must be conside red first of all
in terms of their dyn amics. Bue this dynamic must itself be viewed from
several angles.
T ranee being by definition tr ansitor y, possession must be first ap -
proached from the viewpoint of the dynamics of trance. The modification
32 CHAPT ERO E

of the state of consciousness characteristic of trance follows a process

comprising a sequence of distinct phases whose sequential order obeys a
particular internal logic. The se successi ve phases-preparation, onset,
climax, resolution - each present particularities that va ry significantly
from one cult to ano ther. As for the modalities of this modification-more
or less imperceptible or brut al, mild o r convulsive, calm or dramatic-
their variations are equally significant. The rel ations that trance ma y ha ve
with music can va ry not only from one phz.se to another but also accord-
ing to the type of modification invol ved. Thi s being said, however fleeting
or brief the state of possession ma y be, it is always more or less durable. It
is therefore important to consider it also from this aspect, the one it takes
on during its culminating phase. This phase, which varies in duration and
can, I repe at, be very fleeting, consists in a state of consciousness charac-
terized by a form and a content that are both set and stable. It is a slice of
time experienced under an identity othe r than the one of everyday life, and
which we can justifiabl y isolate from the let us say normal state of con-
sciousness that precedes and follows it. During this phase the adept is no
longer a Niger peasant or a Bahia street vendor but Dong o, go d of light-
ning, or Yemanja , goddess of water. What is significant, then, is the be-
havior of the spirit incarnated within the possessed person.
In additio n, trance, both as an experience and as a corporeal technique,
represents different types of events from the adept's standpoi nt depending
on whether he is integrating this experience and learning this technique-
during initiation - or whe ther he exercises previousl y acqu ired knowledge
and skill. Initiation, which is the discover y of a new state of consciousness
and an apprentice~h ip in the var ious ways of reaching and then leaving it ,
is a unique experience that will not be repeated, and one th at results in an
irreversible modi fication of the person's rel ations w ith himself, with the
di vinit y, and with society. Once this initiation has been undergone, the
modification is permanent. On the other hand, the inner change th at the
vo tar y undergoes when he passes from his ever yda y state to his possessed
state, and then reverts to his former one, naturall y recurs ever y time he
goes into a trance through o ut the remainder of his life. But, and all the de-
scriptions concur on this point, the behavioral patte rns characterizing
trance differ according to wheth er it in vo lves a no vice, a recent initiate, a
confi rm ed adept, or an officiant experienced in the trance state. The be-
havio r of the possessed person thus va ries during his career, and his trance
takes different form s depending on the stage he has reached. It is therefore
essential to set tr ance w ithin this dynamic process w hen one investigates
its relations with music.
Lastly, trance, as an event, is linked with the successive stages of a cere-
mon y, and does not generally occur at just any time. A possession ritual is
an architecture of time also composed of va rious phases connected with
 ~/f '"'°...,_·,
-s,;""'-..::· ,/ ~
~}- ~ -~ c.;
4Jy ~,~~
T ranee and Possession ,1,,, D lNFOR\f\i,:~ 33
~tilDfJL Ul'U'l'IP\
different kinds of music. It is thus within the dynamics of the ceremo ny
that we need to consider the relations between music and trance.
Investigating the relations between music and posses sion, ultimately
means asking ourselves what role music pla ys in the preparation for
trance, iri the onset of trance, in the m aintenance of trance, then in the
emergence from trance, all this being seen in terms of the stage the ade pt
has attained in his career and of the particular moment in the ceremony
that has been re ached. It is within thi s threefold dynamic - of trance itself,
of the ade pr 's life, and of the ritual's development - that we w ill therefore
exam in e these relations.
So what means do we have at our disposal to conduct an investigation
inro these relations between music and possession? To begin, let me ob -
serve that alt h ough the various manifestations of possession have often
been likened - whether rightly or wrongly -to those of hysteria, we do
nor know a great deal about rhe actual st are experienced by those who
ha ve become possessed. Th e descriptions at our disposal have all been
made from the outside, apart from the one by G. Cossard (1970 ), wh ich
despite its qualities, is nevertheless extremely brief. On the othe r hand, al-
though we know that music does ha ve dynamogenic effects, although ex-
periments have shown that certain sounds or certain kinds of music en-
courage the growth of plants, rhe lactation of cows, or the resignation of
saleswomen in department stores, although, in short, we can be sure that
music does have the capacity to act upon the psychological or ph ysiolog i-
cal beha vior of living beings, these effects tell us nothing about the wa y
music may act in the case that concerns us here , which is totally distinct
from all those listed above. The onl y research carried our in this field, rhar
o f 1 eher ( 1962 ), is unfortunatel y de void of all scientific value , as we shall
see later. We therefore ha ve no experimental dat a available for ad vancing
an y hypothesis on the psychophysiological role of music in trance stares.
Given the present stare of things, the only means of understanding this
role is ro turn to the existing evidence we ha ve on possession cults, which
is to sa y, essentially, the writings of ethnographers. These ethnographers
(or anthropologists ) provide us with three kinds of information, depend-
ing on whether the y offer accounts of the correlations natives make be-
tween music and tr a nce - unhappil y th e rarest kind - whether they them-
sel ves make such correlations - unconsciously more often than nor - or
w hether the y simpl y describe the fact s in such a wa y as to enable us to dis-
cern concomit ances between music and trance oursel ves, which is rhe
most frequent kind, though in this last case precise obser vat ions are much
less numerous th an one would wish.
In examining the rel ations between music and possession as the y emerge
from such rexrs, or rather from those I have selected as being particularly
significant, I shall conform to rhe princip le stared earlier, char of consider -

196 0 8
34 CHAPTERO E

ing possession cults as dynamic wholes. I shall therefore proceed by fol-

lowing the various stages the adept goes through in the course of his ·career
as a possessed person, beginning with the stage I shall label as preposses-
sion. But first it is import ant to provide a bro ad outl ine of this career and
of its principal stages.

Possession and Vocation

Whether explicit or not, the aurochthonous theor y is that possession is
produced by the will of a di vinit y who, for one reason or another, wishes
to inc arnate him- or herself within a particular individual. The god is the
one who selects the person in whom he wishes co incarnate and not the re-
verse: a person does not select a god for whom he then becomes the vessel.
On the surface, at any rate, the relation as experienced by the possessed
person is passive rather than actively willed.
In Vietnam, in the hizi'i bong, "it is accepted that Genii themselves are
the ones who select their mediums, 'catch' them, and oblige them, will y
nill y, to serve their masters" (Simon and Simon -Barouh 19 7 3, 33 ). In the
iger, among the H ausa, "societal viewpoint requires that ade pts be cho-
sen by and solely by the gods. Any initiative on the part of the wo man con -
cerned would only be merel y presumptuous and sacrilegious" (Monfouga-
icolas 1972, 95 ). In Dahome y, among the Gun, "the vodun himself
selects vodun si from members of the families who make up his followers
(Akindele and Aguessy 1953, 113 ). In Brazil, among adepts of the can -
domble, "a deity's supreme requirement is that one of the faithful become
his 'horse' by means of trance , so that he can come down among mankind
and dance" (Cossard 197 0, 95 ).
Perhaps we ought co detail the nature of this relationship. The fact that
initiation is sought (as it is in many but not all cases, as we shall see ) does
not, in practice, mean that it has been deliberatel y willed. To be initiated
into the wo rship of a di vinit y when yo u know, thanks to the inter vention
of a soothsayer, that the divinit y has chosen you, constitutes a submission
to th at choice , not an act of choice on your p art. Whether in Bali (Belo
196 0, esp . 2 , 47 ff. ), T anzania (Gray 1969, 175 ), or in Dahome y (Akin-
dele and Aguessy 19 5 3, 114 ), the witchdoctor or soo thsa yer is the one
who, in answer to the questions of a person who has come to express his
or her troubles, un veils behind the client 's difficulties either the anger of an
offended divinity or th at di vini ty's quest for a receptacie, a visible enve-
lope . If the soothsayer's interpretation meshes with the client 's uncon-
scious aspirations, so much the bette r; as with good matrimonial agen -
cies, good soothsayers are those who can arrange alliances sa tisfacto ry to
both parties. "Officially," to use the term emplo yed by Jacqueline Mon-
fouga -Nicolas (1972 , 93 ), he ha s sim ply translated the wishes of the in-
visibl e element that governs the world.
T ranee and Possession 35
Thu s, if one becomes an adept of a possession cult, it is not because one
has willed it oneself - o r at any rate not because one has w illed it con- ·
sciousl y-but because a god wishes it. I should add that it is not always a
soothsayer who discloses this fact, and w h o identifies the go d requesting
the in itiatio n , as is the practice in Ch ad, for ex ample , among the Mun-
dang (Adle r and Zempleni 1972, esp. 87 , 118 ). In a neighboring countr y,
Nige r , among the H ausa (Monfouga - icolas 1972, 97 ), on the contrar y,
"the healer of the bori [the woman in cha rge of the college of possessed
wo men ] is th e one who decides wh ether o r not a wo man had definitely
been chosen by any particu la r god." But this does not affec t the problem in
any way. The sign ificant point is th at in each case divination is involved;
that is to say, a method of acq uiring information about the w ill of the gods
so that their wishes ma y be obeyed.
In all cases, that is , except for tar ant ism , in wh ich re co urs e to any kind
of divina ti on is never described. T his exception, ho wever, is neither com-
plete nor sig nific ant. Ob viously, t ara nt ism is an instituti on in an advance d
state of decadence. Although it still att racts "new recruits" (cf. De Mar -
tino 1966, 41 ), we know not hing about the way in w hich the y are recrui t-
ed . Bur in any event it offers a case of the "selec tion by illne ss" mentioned
by J. Monfouga - icolas (1972, 97 ), a selection process that does in fact
cons titute a system of divination, either in itself or as an element of a
broader divinatory system, as in the case of th e Mundang (Adle r an d
Zempleni, ibid. ). Moreover identif ying the ta r antula responsib le for the
bite, in oth er words the possession, is carried our by means of a "musical
diagnos is" that also functions as a divinatory sys tem. We shall return to
this subject later.
.Alongside possession by personal vocation, if one may so express it,
which is a response to a particular st are or unhappy event in th e life of an
individual (as is the case, among other examples , with the sadeg "trance
medium" in Bali, and w ith oth er categories of possession : zZir in Ethiopi a;
ra r anti sm in Europe; the candomble in Brazil; and usually wit h bori
adepts ), there also exists a kind of possession that affects the person un -
dergo in g it not as an individual bur as a member of a group. Two ways are
possible here. In the less frequent of the two, possession is handed do w n
as an inheritance, wh ich can occur in the case of bori: "Bori can be trans -
mitted as an inhe r itance, in extremely r are cases, without the intervention
of any pathological elements" ( Monfouga -Nicolas 1972, 93 ). The same
thing has been obse r ved in Bali, bur here aga in , as an except ion (Belo
1960, 19- 20 ). Among the Fon of D aho me y, Her skovits tell s us (1938,
2: 19 5, 197 ), albeit wi th ou t clarifying whe ther or not the practice is com-
mon, the adept of a vodun can designate his successo r before dying. In the
othe r way the family, group, o r village consecrates an individual to the
gods, usually at a very early age. T h is was the case in Bali -J ane Belo
(1960, 4 ) speaks of it in the past tense - w ith girl da ncer s, under the age of
36 CHAP TE R O E

puberty, whose trances owed

•
nothing to illness. These girls w ere in pr ac-
tice chosen by the village whe n it decided to form a "cl ub " around two fe-
male dancers in the hope of dri ving out epidemics. Little girls , Belo tell s'
us, "were chosen appa rently for their ability to go into the disassociated
st ate and to perfo rm in this sta te dances which it was said th ey co uld not
possibly have executed if they were not entered by a god, since the y had
received no previous tr ai ning in the dance." This is a lso most frequentl y
the case in vodun cults among the Gun of Dahomey, at least when a lin-
eage vodun is concerned. Since filiation among the Gun is patrilineal, th e
child is consecrated as a general rule to the vodun of the fa ther's lineage,
on the adv ice of- or at any rate afte r consultation w ith- t he soot hsaye r.
The recruitment of vodunsi, the "w ives of the vodun, '"" thus follows a
path opposite to that of the borz or zar. In the case of the vodun it is a fam-
ily matter, and the indiv id ual concerned (usually a girl ) is not even con -
sulted (a situation comparable to the marriage process prevalent among
the Gun, in its traditional form at least ); in the case of zar or bori, on the
contrary, it is above all an indi vidual matter, 75 am ! althoug h it may in vo lve
familial components, these are not decisive. Thi s does not mean that in
the case of vodun personal predisposition does nor play a certain role. In
order to choose future vodunsi, and thu s prepare for the next "class" of
adep ts , "several months before the ceremony of their entry into the con -
vent, " the kinfolk, with the agreement of the priest of the vodun con -
cerned, consult Fa , the god of divinat ion . The list of "pupi ls selected by
the vo dun " is established in this way (Akindele and Aguessy 1953 , 114 ).
Among other things, this consultation obviously bears upon a particular
child's predispositions w it h regard to th e status of a vodun's wife, and th e
final choice certainly t akes this into account. The fac t remains, neverthe -
less, that the decision ultimate ly rests with the family. Vodun and lineage
live in a reciprocal relationship, since the more powe rful the vodun the
greater the protection for the lineage, and the more the lineage hono rs its
vodun (notably through la\·ish ceremonies, and therefore by large num-
bers of faithful followe rs) the m ore the deity's power is consol id ated. Thi s
being so, it follows that the lineage must ensure its bonds with the vodun
do not become atten uated in any way . It mu t therefore, regularl y suppl y
its vodun, if one may so express it, w ith ne,\· wives as the generations age
and rene\\· themselves. It can happen, however, that the family's zeal
slackens, or, as is becoming more and more frequent , that new recruit s,
influenced bY marabouts and missionari r s of va rious denominations, are
recalcitrant, refuse to enter the convent \\·hen the moment arrives, or even
run away from it. Any misfortune befalling the family or the country is
then interprete d as a wa rning to restore order. We know that the 197 3
drought was regarded by a number of Africans as a sign of the gods' anger
at being deserted by m ankind. Attempts were consequently made to revive
T ranee and Possession 37
old form s of worsh ip that had eithe r become extinct or we re on the verge
of becoming so (de Sardan 1973; see filmograph y ). Possess ion is thus
so metimes a given group's response to adversi ty, a situa tion analogous to
the o ne we already mentioned, in which it constitutes an individual re-
sponse to illness.
Vocation of the impersonal type, by wh ich I mean that in which the
god's choice (since, let us recall, it is always the god who does the choosing
ultima tel y ) affec ts the person involved (usuall y a child ) not as an ind ivid -
ual but as the rep resentat ive of a lineage, can take othe r forms as we ll. It
often happens th at if a woman rem ains sterile, or regularly loses her chil -
dren ve ry young, for example, her parents decide to consecrate any future
ch ild in advance to the vodun from whom, always on the advice of th e
soo ths ay er, assista nce will be sought . 76 Th e vodun in ques ti on will then be
one of those that Akindele and Aguessy term "private voduns, " particular-
ly D an, the snake god, or Sakpata, the earth god, who controls the inci-
dence of smallpox. It sho uld also be observed th at althoug h he ma y be a
"p ri vate vodun" for certain families, Sakpara (one of the most po pul ar of
all vodun in the Gun regi o n ) may equally be a lineage vodun for othe r
families.
The orisha cult among the Yoruba, neighbors of the Gun, has t wo types
of recruitment process, one personal and one imperson a l. In the ago ter -
ritor y" of Benin , Shango , who is the dominant orisha , recru its his vota -
ries yo ung, some tim es as very yo ung children, I am told by Pierre Verger ,
who once saw a novice of about two yea rs old still be in g ca rried on his
mother 's back. It is clea r that such recruitment, classifi ed by age, and in -
tended to renew the all iance between fam ily and divinit y as we ll as to en -
,ture the succession between gene r ations, rakes place at the family level,
w ithout consulting with the individual selected. In short, it fa lls into the
catego ry of wha t I have termed impersonal voca tion, because it is unc on-
nected with the indi vidu al's personal histor y. Pierre Verger observes
(1969, 51 ) th at the most frequent form of recruitment occurs at the mo-
ment of birth, when the oracle indicates that "such is the child's dest in y."
But it can also happen, and app arent ly thi s is not exceptional, that t he ori -
sha acqu ires recruits the other way . \erger ( 1957, 159 ) recounts the case
of a woman whose children had all been stillborn, victims of an evil sp iri t.
Ogun, god of iron, the village's predominant orisha, then took this wom -
an unde r his protection (afte r she had lost consciousness at the door of his
temple ); she underwent initiation, and then became an Ogun vo tar y.
Her e, the recruitment is of the same type as the one that predominates in
the bori, ziir, and ndbp cults. 78
The vodun and orisha cults share one feature that distinguishes them
from a whole group of other interrelated cults : the y are both, not at the
collective but at the individual level, strictly monotheistic. The vo tary is,
l•

38 CHAPTER ONE

indeed, consecr ated to one and onl y one vod un or o rish a, ' 9 who in the
great majority of cases is the adept's lineage deity. To use the term em-
plo yed by Luc de Heu sch ( 1971, 254 ), possession in this case is the deed of
an endogenic spirit. In bori , as in zZir or hdu bong , the votar y is possessed
by several spirit s in succession. In bo ri it would seem that these spirits are
mostl y unconnected with the votar y's lineage, in other words the y are al-
logenic. In zZir the spirits are most often seen as coming from another dis -
trict, and therefore are once again allogenic; but th.is rule is not absolute. 80
Let us remember in passing that Jeanne des Anges of Loudun was pos-
sessed by seven demons (Certeau 1970, 136 ). In tarantism, on the con-
trar y, the person affected is associated with one particular species of taran-
tul a and no other , a fact that results in hi s selective sensitivit y to a
particular form of the tarantella. On the other hand, the vodun and orisha
cults are distinguished by the fact that in the first, initiation entails ap-
prenticeship in a series of secret langu ages ,81 whereas in the second, Pierre
Verger tells me , there is no such component.

Trance , Crisis , Fit, "Wild" Possession

Three w ords constantl y recur in texts describing possession cults: posses -
sion (of course ), trance, and crisis (crise )." But all authors do not use the
last tw o terms in the same way . Some talk m ainl y of crisis: Michel Leiris
(1958 ) and Andras Zempleni (1966 ) in particular , in their descriptions of
possessi on in Ethiopi a and Senegal respecti vely. Other writers mostl y use
the w ord trance and emplo y crise solel y with reference to one of its par-
ticular aspect s. This is true for Jacqueline Mofounga-Nicolas (1972 ),
among o ther s, and limiting oursel ves to texts in French. Other writers use
both word s almost interch ange ably. The w ord crise is used to denote
sometimes a con vulsi ve form of trance, sometime s the parox ysmic aspect
o f trance w hen it displ ay s peri ods o f va rial:Jle intensit y, sometimes the on-
set of tr ance if it is ch aracter ized by a moment of more or less distressing
agit ati on , and finall y sometimes - and this last is a very different usage-
to re fer to the troubled peri od a given indi vidu al m ay undergo , and which
will lead him to seek an o utlet in possessi on. It is in this last sense that
H. A. Junod (1913 , 2 :43 8), among others , uses the word: "Thus the first
signs of a posses sion seem to be a ner vous crisis , but also the occurrence of
certain suspici ous symptom s . ... However these symptoms are not suffi -
cient ground fo r a di agnosis , and the bones are alw ays consulted to come
to a conclusion, " he w rites in his fam ous monogr aph on the South African
Thong a . But several pages later Jun o d uses the w ord in another sense, to
denote the onset o f possessi on: "Conquered at length by this nois y con-
cert, the patient enters a condition of ner vous exaltation. The crisis oc-
curs, the result of evident hypnotic suggestion. He rises and dances wild -
ly ... " (ibid. , 443 ).
Trance and Possession 39
E. De Martino, in his work on tar antism, also uses crise in nvo different
senses: "The real therapeutic ritual begins with music. At first there was
solely the crisis, that is to say, the culturally conditioned imitation done by
the poisoned victim . . .. Within the framework of this symbolic imita-
tion, the tarantulees refer to themselves as feeling 'broken,' 'to m, '
'bruised,' 'shattered,' or else- as I have so often heard said during our
fieldwork - 'wounded,' or 'upset,' in other words filled with a profound
dist aste for oneself and for everyt hin g. The extreme form of this imitation
is an ab rupt collapse, which means that the rarantulee is enacting the role
of a person in danger of dying ... " (1966, 142 ). On the preceding page,
the wo rd crisis is used to denote a quite different stage, the one in which
the tarantulees, after having already danced a long time, meet together in
the chapel of Saint Paul and communally attain the parox ysm of their
trance , a parox ysm Martino describes elsewhere in these wo rds : "in the
interweaving of the indi vidual crises .. .the gener al and desperate agita-
tion was domin at ed by the stylized cry of the tarantulees, the 'crisis cry,' an
ahiii uttered w ith va rious modul ations, that sounds more like the yelping
of a dog than a human cry" (ibid ., 121 ).
A. Metraux, in his stud y of H aitian voodoo, reserves the wo rd crise for
the initial stage of possession, in which "the trance manifests itself by
symptoms of a clearl y psychopathological character " (1958, 106 ). Roger
Bastide , on the contr ary, sometimes uses crise and transe interch angeabl y,
so that in his series of studies on tr ance states we can re ad , only a few pages
apart, "the crise brought on by the bossal god" and "trance brought on by
the loa bossal" (1972, 88, 92 ). Elsew here (on p. 74 ), he write s of posses-
sion by the bossal god. 83 Luc de Heu sch (1971, 228 ) uses the terms transe
and crise, or even crise de possession and crise extatique w ithout differen-
tiation. For our purp oses, however, it is indispensib le to assign a clear
meaning to each of thes e wo rds , or, rather, to make it quite clear to which
precise ph ase of possession the y refer. Further , it is essential to realize that
the var ious possession cults throughout the world differ greatly in the type
of temporal development the y exhibit. Some go through numerous, very
ritu alized stages, the milestones of a long initiation; others, on the con -
tr ary- and tarantism in particular-display a very telescoped type of de-
velop ment . Certain distinctions that are thus indispensib le when describ -
ing certain cults ha ve no relevance for others.
Fir st , let it be said th at if crisis and trance are so often used interchange-
ably it is because they both de signate an abnormal and tr ansito ry state.
But alth ough tr ance is, in th is sense, always a crisis or a fit, crisis, on the
other hand , is not always a tr ance. loreo ver, viewed as a temporal devel-
opment, trance contains a particul ar stage -entr y int o the trance - which
itself often rakes the form of a crisis and , consequentl y of a crisis within a
crisis. Finall y, as we have seen, the wo rd crisis is also used to designate the
state of crisis, either chro nic or acute , that leads the individual who expe -
40 CHAPTER ONE

rienc es it to enter the path of possession. It is th ere fore important to clari fy

the amb iguity inherent in this term.
The chronic crisis to wh ich I have just referr ed -illn ess, misfortune ,
troubles of all kin ds aff ectin g the indi vidual 's person, human relati on-
ship s, or posse ssio ns- leads in general co an acute nervo us fit interpreted
as th e sign that a divinit y desires embodiment. Who chis divinit y is rem ains
to be disco vered. If it has already been identified, an all iance w ith it re -
ma ins to be made . This ner vous fit often consists in falling unconscious to
th e grou nd . Frequentl y, this event occ ur s w ith o ut any kind of cere mon y, as
in tarantism, w here , as we ha ve just seen, this sudden collapse repre sents
the parox ysm of the fit. Describing an initi at ion in the Yoruba region of
D ahomey, P. Verger ( 19 57, 159 ) rel aces how the wo man in question had
pre viousl y "fallen co the grou nd in a rigid scare" o utside th e temple of
Ogun , god of iron . In Chad, among the Gul a Iro, all possessed women
"scare their ca reer" as the result of a fie of thi s sore char floo rs them quite
unexpectedly (Pairault 1969, 25 8). In T anz an ia, amo ng the Segeju (G ray
1969, 175 ), the victi m is brut ally thro w n co the ground "as if by some in-
vincible force." The da y an d precise rime of the fall are used by the diviner
co identif y the shetani geni us responsible for ic. Bur the fall ma y equall y oc -
cur during a cerem o ny. In Benin thi s is usuall y the case in the ritual s involv -
ing Sakpaca, god of the earth and master of smallpox . T his also hold s true
for the candomble (Cossard 1970, 15 8- 59 ). bur Cossard specifies (ibid . ,
162 ) chat the future adept "may also fall at anv oche r given moment out -
side the ceremonies."
I would like to propose w e label this first son of crisis, which resem ble s
a ner vo us fir or att ack much more than a possess ion cris is, a "prepos ses -
sion" cr isis o r fie. Whether or nor it rakes the form of a fall, such a crisis
co nstitute s in man y cults a unique a nd dec isive type of eve nt in th e life of
the future adept. In man y cases, th e cr isis (and fall ) does nor merel y signif y
that the divinit y has taken possession of the indiv idu al, it also ind ica tes,
a bo ve all, that the individual has been "k illed " by the divin ity. The fall oc-
cupies a p articul arly import ant pl ac e in the ndbp cult amo ng the Lebu and
the Wolof of Senegal. Whether a "no nriruali zed pos sessio n crisis" (in oche r
words a "prep oss ession" crisi s) or a "ritu al cri sis"' is involved, the "fall"
cons titutes the "termin al event," the "natu ral conclusio n" of the crisis
(Zempleni 1966, 314 , 40 0,4 10,414 ). As the resol uti on of a ritu al cri sis,
it of course occu rs during the ceremony. As the resoluti o n of a nonritual -
ized cris is, o n the other hand , it m ay a lso occur - like the crisis itself-out-
side any ceremon y, usually "at home. "•· In rhe ndop the fall does not there-
fo re constitute a unique event but rathe r a recur rent one . evenheless we
sh o uld note th at those ad ept s wh o p reside ove r the cu lt, the "offic iant s" or
boroom rab bu mag (ibid. , 3 06 ) fall only very rarely.
In the case of tarantism, as with ndbp, the fall ex hibits two disti nct
cha r acteri stics. As we have seen, it occu rs on the one hand as a sign of pre-
T ranee and Possession 41
possession, and on the other as the end product of the dance itself (D e
Martino 1966, 69 and photo 7 ). In the case of the Balinese barong, the
possessed dancers, h av ing atta ined the parox ysm of excitation during
their kriss dance, then fall to the ground (Belo 1960 , 98 ). In Tibet, medi-
ums fall unconscious after the y h ave delivered their oracles.
Zempleni ( 1966, 419 ) interprets this fall in the case of the ndbp as "the
fatal outcome of the meeting with the double," which is, moreover, what
the "symb olic burial" episode illustrates.
Among the Gun, the enclosure where the novices live during their reclu-
sion - an enclosure situated within the boundaries of the vodun's tem-
ple - is some times referred to as kuxwe or "ho use of death" (Tardies 1958,
53 ). Several days after rhe entry of a graduating class of recruits into the
convent, their parents thank the vodun's priest for "hav ing restored their
children to life " (Akindele and Aguessy 1953 , 114 ), a fact clearly demon -
strating that rhe children, upon entering their new state, have undergone a
symbolic death. It would seem that their "death" and subsequent resurrec-
tion occu r our of sight, within the secrecy of the convent. In fact, the ritual
always differs to a certain degree depending on the vod un concerned. In
rhe case of Sakpata, the recruits ' death and resurre ction are public and
rake on a particularl y dramatic and spectacula r character. 85 But this is the
exception, not the rule, contrar y tO what is usuall y assu med.
The amalgam of possession crisis (o r rather prepossession crisis ) and
ritual death is also found among the Yoruba , as is implied by the following
phr ase s extracted from an orik i (sequence of praise mottoes ) of Shango,
go d of lightning (Rouget 1965, 102- 3):
Th e earrh does nor ear for ever
Leopard-Dearh , Dearh rhar sends rhe child ro sleep'
Death rhar bruises many rimes before raking a \\·ife1

Behind these esoteric phrases one should read that the ritu al death is mere-
ly temporary ("not ... for ever " ) and that Shango , who is Death insofar
as he kills ("sends the child tO sleep" ) and also Leopard insofar as he is
king, su bjects the no vice tO a long ordeal ("bruises man y rimes ") during the
reclusion before finally making him or her his adept ("his wife" ). This text
poses the problem of the relations between ritual death (however realisti-
cally ac ted out in the case of Sakpata ) and the first possession crisis. How
are we to accept that the recruit can be at the same time dead, since the god
has just killed him or her, and also possessed by him - since our authors in -
dicate, in the case of both vodu n and orisha, that this first crisis is a posses -
sion? This contradiction might perhaps be resolved if w e viewed the mat-
ter from the standpoint of native th ought, which believes that human
beings possess several souls. In Metraux, for instan ce (1958 , 106 ), we
read that "the explanation given by members of the voodoo sect for m ystic
trance is among the simplest: a loa takes up residence in the head of an in-
42 CHAPTER O E

di vi dual after ha ving first driven out the "big good a ngel ," one of the two
so uls that ever yo ne carries within him. The abrupt departure of the soul
ca uses the twitching and violem movemems th at characterize the onset of
trance. Once the "goo d angel" ha s left the po ssess ed pers o n exper iences a
feeling of total emptiness, as if he were losing conscio usn ess . " 86 We know
that initi a tion into the cult of a vodu n or an orisha is a long process in-
tended to produce profound modific ati o ns in th e indi vidual 's personalit y
structure, so that he or she wi ll become a recept acl e fit for the di vinit y. It is
characterized by a seq uence of phases during w hich the no vice's he ad is
sub jected to a va riet y of tre a tment s : in candomble, the bori ritu als; in-
tended to "s trengthen the n ov ice's he ad" ( Cassard 1970, 17 6 ); in H a itian
voodoo, the "headwas h" (/aver -tete; Metraux 1958, 107 ); in the orisha
cult the imposition of osu , a mystic preparation placed on the no vice's
shaved head at a cert ai n stage in the initi ati on in o rder to ai d reception of
the orisha (Ve rger 1957, 71 ).
The structure of possession t hus appears to be related to th at of the per-
son, in as much as the lat ter is for med by the conjunction of seve ral so uls.
Perhaps it is in these term s that it can best be described . Maybe its differ -
ent phases could be sa ti sfa cto ril y defined in every case if we could always
refer them to an indigenous theor y of the conjunction and disjunction of
souls . T his is w h at we are led to believe, for example, by wha t Adler a nd
Zempleni ( 1972, 130 - 31 ) tell us about the Mundang , amo ng w hom po s-
session gives rise to a spect acula r "co ming a nd go ing of the souls of the
sinri" " between the sick person a nd the ther apist. During thi s co ming a nd
go in g, the y note the patient is not guided by hi s o r her own "s m a ll soul"
but by the so ul s of the sinri who possess him or her, a nd one hears dia -
logues such as the followi ng between th e patient a nd his or her initi a ted
sister who asks : "H as yo ur little so ul come back, - Yes, it has come b ack ."
Among the Songha y, J. R o uch wri tes ( 1973, 531 ), "th e most w idespread
zima 89 theory is that during possession the "double" (bia )9 0 of the god has
taken the place of the horse's "dou b le." El sewhe re, the Dogo n themselves
describe the mechanism of the cri sis undergone by a candidate for the
function of totemic prie st as the direct re sult of the cont radictor y mo ve -
ment s of his different so ul s (kikunu ). "The struggle between a ntagonistic
kikunu in this crisis symbolizes the struggle between re ason an d affectivi-
t y, wh ich is the source of d isequili b rium ," 9 1 Gene\ ·ieve Calame-Griaule
tells us ( 1965, 46 ).
Let us now return to sym bolic death a nd to the vodun . During their pe-
riod of reclus ion, the vodun 's ne\,. recruit s are in a state of partial death,
since, a lth ough the y rem ai n quite a live all the w hile, as each one can see ,
the y ne ve rtheless displa y behavior associa te d with death , such as eating
\\' ith th eir lef t hands, for example, or go ing through doorways b ack -
wards. In borz, the initiation is described bv J. Monfouga -Nicolas as com-
T ranee and Possession 43
posed of two successive stages, the first an exorcism ,9' the second an im-
position. The former is accompanied by symbolic death: the future a dept
lies fla t on the ground , and this is referred tO as "lying w ith the worms," 93
an expression evoking an interred body eaten by wo rm s (Monfouga-
Nicolas 1972, 117 ). But this death occupies no more than a very secon-
dar y place in the ritual. The "struggle with the dark gods, " on the con -
trar y, which te rmin ates this first phase, pla ys an essential role and rakes a
very SiJecracular fo rm . We might be justified in seeing this as a form of
w h a t R. Bastide (1972, 92 ) has termed "wild trance ," which leads us co
di stinguish it from the prepossession crisis.
Unlike prepossession crisis, "wild" possession cr isis a s described in Da -
home~ ·, Br az il, and Haiti occur s when initi a tion h as already begun.
Among the Fon, Hersko vits (1 938, 186 - 87 ) describes it as a "war " w ith
the vodun , who suddenl y att ack s the new recruits .9' In Haiti, wild posses-
sio n is ca used b y a loa bossa l according co Merraux , who in this context
o bser ves that "the n ature of the ner vous attack depends . . . on the ritu al
co ndi tio n o f the po sses sed person; rhe less experienced he is rhe mo re he
stru ggl es . U nt il hi s he a d has been 'w ashed, ' thar is t0 sa y, as long a s h is loa
h a s no t been fixed in his he ad by a speci a l ce remon y, he behaves wildly"
( 195 8, 107 - 8 ). In rhe cando m ble ir seems that rhe po ssession cris is by a
santo bru to or "w ild saint '" ma y correspo n d either to what I la bel a prepos-
session cri sis- a s Ba stide 's ( 19 5 8 , 22 ) descri ption suggests - o r to wha r
Bas rid e himself refer s to elsewhere (1972, 92 ) as wild possession, in other
wo rd p oss essi o n char h as nor yet been "socialized ." This posse ssio n by a
santo bruto does not necessaril y rake a brut al form, however. G. Cossard ,
in effecr , simpl y likens the sant o brut o ta an orisha who h as not been "edu -
ca ted , do es not know how tO beha ve .. . and still h a s e\ ·er ythin g to learn."
O n the ocher hand , she al so describes (ibid. , 204 ) a ph ase o f the initiation
r iru al , w hich has now fallen into disuse, th a t ver y clos ely resembles rhe
o ne de scribed by Hersko vits to \, ·hich I ju st alluded. This would impl y
there is yet a nother aspect of "wild" possessi o n in rhe candomble . In an y
cas e, the imp o rtant thing to remember is th a t wild po ssessi on , wh a te ver
fo rm it takes , corresponds to a period during which the adept is nor yet
co mpletel y initi a ted.
Imp o rt a nt as it ma y be to di stinguish bet w een prep os ses sio n crisi s a nd
\\·ild poss es sio n crisi s, it is ju st a s important to differenti a te bet w een wild
possess io n a nd th e \,·ildne ss o f p oss ession . 95 We have a l re a d\ · seen th at
posses io n b v a santo bru to in the candomble is not nece ssa rih · brut a l. Ex-
rr emeh · vio lent a nd brut a l p os sessio n trance s , such as cho se filmed in J.
Rou ch 's Les i\I aitre s Jau s- possessions in volving such truh · sa vage forms
o f self -mutilati o n a s can be observed in the Ivor y Co a st o r Tog o- are not
in an y w ay ex a mples of pos sessions of adepts undergoing initiation. By
their ver y nature, such trances demand frenz y. Th e same is true of Balinese
 44 CHAPTER ONE

possessions duri ng which men in trance states turn their krisses aga inst
themselves. But we must not fo rget there also are non violent - even
calm - trances, and o ther s- as is ofte n the case in the Dahome y orish~
and vodun cults - in wh ich possession is so interiorized as to go almost
unperceived by an inexperienced spectator. To take two cults studied by
Andras Zempleni as examples: possess ion among the Wolof of Senegal is
as brutal and convulsive as tha t amo ng the Mundang of Chad is gentle and
relaxed. As two films demonstrate 96 the behavior of the possessed is total-
ly different in each of the two cases. Th e music involved is just as different ,
a sign ificant point to w hich I shall return .
Finall y, as far as making a general disLincti on between trance and crisis
or fit is concerned, pe r haps it might be best tO reserve the word "tr ance"
for the mo re or less durable per iod dur ing which the adept full y identifies
himself with the god pos sessi ng him, and the word "crisis" or "f it" fo r the
ver y temporar y, often painful and more o r less convulsive state ma r king
the tr ansition from a norm al to a trance state, or , on the contrary, as in
nd op, the end product of that state.
We would, in this way, be led ultim at ely to dist inguish between trances
with and without crises o r fits; crises with and w ithout fal ls; and , recipro -
cally, between crises or fits and falls with and wi thout trance .

Pos sess ion and Ob sess ion

Befo re b roach ing the subject of initiation, w hose goal is to help the adept
enter into ritualized pos sessio n , or possession in the full sense of this term ,
let us return for a moment tO the state th at is an antechambe r , as it we re,
of true possession and w hich I proposed we call prepossession . We have
seen th at this latter state is frequently characterized by what Zempleni
calls "non -r itualized possession crisis." Clearly, to speak of prepossession
is to impl y that this state precedes another, possession itself, and by the
sa me token, to speak of "nonritualized possess ion " would be meaningless
except with reference to "ri tualized possession," which it w ill become
sooner or later. Frequentl y, however, states corresponding to the picture
just given of this prepossession - or, if one prefers, this nonritualized pos -
session - a re not subsequently tr ansformed, through initiation, into a
state of full possession . \"X/hatterm should we consequently use tO desig-
nate such states?
During th e Renaissance - and without d o ubt long befo re that - a dis -
tinction was made between two states , "obsess ion" and "possession," the
first being regarded as the effect "of th e Demon as an external agent," the
second as that of the Demon "as an internal principle." 9 ; In the case of ob -
sess ion the Demon was thought to act upon obsessed indi viduals "by ap -
pearing to them ofte n and visib ly so, regardless of their w ill in the matter ,

i.....
Trance and Possession 45
by strik ing them, by troubling them, and by exciting in them strange pas-
sions and movements far exceeding the scope of their capacities," whereas
in the case of possession the Demon "disposes of the faculties and organs
of the person possessed in order to produce, not only in her but also
through her, actions that the person could not produce by herself." In
Loudun - for these quotat ions refer to the Loudun possessions - it does
indeed seem that only those nuns whose demons had been identified were
seen as possessed; the others, whose "sym ptoms" were "not so clear " we re
viewed as merely obsessed. In ecclesiastical language, obsession is "the
state of a person thought to be disturbed, besieged by the devil, which is
different from possession, which signifies actual habitation by the devil
wi thin a body. " 9 '
Leaving aside its Christian context and its reference to the Demon , this
obsession I possession distinction corresponds exactly to the one I am try -
ing to make. Nothing is more common throughout the world than the be-
lief th at certain troubles occurring in peoples ' lives, certain bizarre forms
of beh av ior, certain illnesses, particularl y th ose of a mental o r emotional
kind, are to be attributed to the spite of malevolent spirits, baneful fo rces ,
or wrathful divinities, who torment such victims with their presence and
affl ict them body and soul. Everything that happens to such a person is
then interpreted by him as being brought about in such a way. From our
present perspective, we could say that obsess ion refer s precisely to suc h a
sta te. There are two possible ways to put an end to this condition, wh ich
is always experienced as a disorder: either by transforming and organ izing
the relationship with the spirit or di vinit y seen as responsible for the dis-
turbance - wh ich entails recourse to possession ritu als- o r else, by bring -
ing it to an end-which entails recourse to exorcism. Naturally, however,
these alternatives leave room for a number of intermediar y solutions.
In Benin, when one sa ys in Fon vodim wata eton mf!, "the vodun has
come into his head," one is speaking of true possession in rhe sense we
ha ve agreed to assign this word. When, on the contrar y, one says vodun
wa nu x¢, "the vodun has tormented him ," 99 one is speaking of the state of
prepossessio n just described. As Fon clearly expresses it, this latter stare is
conceived and experienced in a completel y different way from the first. To
designate it in English or in French , it wo uld therefore also be preferable
to use a completely different word. Why not "obsession," since this is pre-
cisely the meaning the wo rd once had? This wo uld certainly make for
greater clarit y.
Moreover, if we further agreed to apply the term "o bsession " not only
to prepossession states bur also to those gene rally labeled possession bur
that call for exorcism and not initi ation as a response, we would be spa red
the opposition between "happy" and "unhappy" possession - or "authen-
tic" and "in authent ic"- proposed by Luc de Heusch (1 971, 232 ) as a
46 CHAPTER ON E

means of distinguishing between accepted and welco med possession and

possession that is rejected and treated by exorcism, an op position that has .
the disadvantage of introducing into the problem a value judgment that
can lead to all kinds of confusion.
It ma y be objected that in this cas e the Loudun pos sessions, since they
called for exorcism as a response , should not be regarded as possessions,
even though I ha ve already accepted them as such. Does this mean I am
contradicting myself? A few wo rd s w ill sho w I am not. Loudun in fact il-
lustrates a situation in which both identific ato ry possession and exorcism
were invo lved ; but here the exorcism was of a very particul ar nature, in
that the victim s were not the ones who asked for it ; society (or the Church,
w hich comes to the same thing ) imp osed it. Contrary co w hat happens in a
societ y that proposes possession as a remed y for people 's ills, at Loudun
society entirel y rejected possession, and it was the indi viduals concerned -
who resorted to it, thereb y conforming co an illicit model. Hence exor-
cism, which was repress ive instead of curative , and wh ich did not tr y to re-
mo ve a tension between the visible and the invisible, but to resol ve a con-
flict between the indi vidual and society. Abo ve, I proposed we term the
Loudun possession reprobate. With the exception of reprobate po sses-
sion, then -in which case this qu alific ation must be specified - we w ill
consider that the difference between obsession and possession lies in the
fact that obsession calls for either initiation - thereb y becoming posses-
sion-or exorcism, w hereas possession, which can only occur after initi-
ation (the latter being eventu ally reduced co mere identification of the one
responsible for the trance , as in tarantism ), never in any case leads to exor -
cism.

I 1 1TIATORY T RANCE AND M USIC

Initia tion, w hose duration and comple xity vary a great deal from one cult
to another, is in mo st case s the moment at which the ade pt undergoes his
or her app rentic eship in possession, a nd at w hich music and trance estab -
lish wi thin him their tightest relationship. Consequently, it is of the great -
est int erest for us here. But the very wo rd initiation almos t ine vitabl y im -
plies secrecy. Thi s is why we ar e, as a rule , badly informed about it. The
mos t notable exception undoubtedl y is the Brazilian candomble. Initi-
ation into this cu lt has in fact been remarkabl y we ll described by Gisele
Binon -Cossard (1970, 157 - 217 ) in a stud y' 00 rendered all the more fasci-
nating because it was conducted from the inside, as it we re , since the au-
thor was herself an ade pt. I shall th erefo re begin this chapter w ith initi-
ation to the candomble . I w ill then deal with the orisha cult in Africa , the
candomble's continent of origin , followed by the vod.un cult, w hich is geo -
graphically and religiousl y very close to that of the orisha and has similar -
T ranee and Possession 47
ly spread abroad, to Brazil, Cuba, H aiti, and var ious islands in the An-
tilles .

Initiation, Possess ion, and Dispossess ion in the Brazilian Candomble

and the African Orisha Cult.
In the candombl e, as described by Gisele Binon -Cossard, initiation lasts in
theor y for a period of three months, but the most important stage co nsists
of a three -w eek period of confinement or reclusion. During this entire pe-
riod no vices live in a state of consciousness completel y different from the
one usu ally considered normal; this stare is characterized by two quire dis -
t inct aspects, one termed the santo state , the other the ere state.
The state of sant o (saint ) is experienced when the or isha is embodied, in
o ther w ords w hen the no vice is effecti vely possessed by the divinit y. H ow-
ever , Cassard w rites: "this state ne ver lasts more than a few hours; it is ex -
tremel y tiring ph ysicall y and , in addition, since swa llowing scarcel y oc -
cur s in the course of it , it is impossible for the no vice to rake nouri shment;
consequentl y it cannot be continued for long w ith impunit y. It is therefore
nec essar y to put the no vice into a different state called the state of ere"
(197 0, 167 ). This latter, we are told, is a "less violent state of trance "
(ibid., 164 ). Even though the alterations of consciousness are much less
m arked than during the sant o state, the ere state is still accompan ied by
ph ychophysiolog ical disturbances: extremel y reduced faculties of taste
and smell, a diminished sense of touch, almost no need for sleep, and dis-
sociation from affecti vity relating to normal life. In order to provoke this
the ere has to be summoned , for the ere' 0 ' is in fact a child whom "the ori -
sha leav es behind " when it lea ves. As a result, the novice in a state of ere
beh av es in a w holl y infantile manner. "Her intellectual acti vity ne ver ex-
ceeds that of a ver y young child," Cass ard tell s us. "The initiates who look
after them pro vide them w ith small pieces of wood , rags , and rib bo n s,
that then become dolls, boats, or kites" (ibid. , 165 ). At the same time the
no vice experiences a sort of "splitting of the self " such that the "ere is
awa re of all the no vice's preoccupations when in a normal state but en vis-
ages them w ith total detachment , as though he w ere dealing w ith a com -
plete str anger. He refers to her in speech as "m y daughter " (ibid., 164 ). We
can therefore say that the no vice exists in a state of double consciousness.
Bur this st ate is also a trance , since, apart from the di sturbances alread y
menti o ned , it also gives rise in cert ain case s to "phen o men a th at fall into
the pro vince of p araps ycholog y : projection into sp ace , projection into
pas t and future time, hence, prophec y, prediction, divination. " During the
va rious initiation ceremonie s , Ca ssard writes, "the littl e mother summons
the orisha by calling our the ritu al w ords of greeting specific to that par -
ticul ar spirit , and by sh a king the ad za. ' 0 ' The stat e o f e.-e then gives w ay
48 CHAPTER ONE

once aga in to the santo state. Afte r the ceremony, the ere state is induced
aga in as described earlier" (by summ on ing the ere three times in a loud
voice ).
In his stud y of the candombl e in Bahi a (w hich differs slightl y from the
one in Rio de Janeiro described by Cossard ), Bastide (1958 ) investigates
the ere sta te ' 03 at great length in his chapter entitled "The Structure of Ec-
stasy." "The crisis brought on by the ere, " he sum s up elsewhere (197 2, 8 8)
"is of two kinds: imit ation of childlike behavior (after initi at ion ) or else a
semitrance (afte r total tr ance )." Hersko vits (1943, 505 ), 1 0 • writing about
yet another region of Brazil, tells us th at "when the possessed person re-
turns to her normal state she must pass through the ere condition, a 'type
of semi-possessi on' described as a childlike state that acco mp anies the per-
sonal diviniry." Thus the ere state is described by Cossard as a "less violent
trance, " by Bastide as "sem i-trance ," and by Her skov its as "semi-posses-
sion," all three term s being slightl y vague and, it must be sa id , somewhat
un satisfactor y. Although closel y linked with possession trance, since it al-
tern ates w ith it through out the initiation period, the ere tr anc e is clearly
distinct from th at of pos session proper, and even in opp osition to it. What
ultim ately characterizes the ere tr anc e from the subject's viewpoint is that
he or she is no longer po ssessed but at the same time has not yet returned to
his or her normal identit y. Srricrly spe aking, thi s involves a double dispos-
session , since the subject is disp ossessed both of the god and of himself . It
may be objec ted th at he is said to be po ssessed by a child. This objection
cannot be sustained, however, since apart fro m the fact that this pos ses-
sion ma y well be purely metaphorical (the texts are not abso lutel y clear on
this point ), identification with a chi ld in these circumsta nces does indeed
co nstitute a dispossess ion, since the child in ques tion is anony mous . Pos -
sessio n proper is the diametric opposite of anon ymit y. Indeed, one is pos -
sessed by a precise deit y who has a well -kno wn name. In the ere stat e , on
the contrary, it is not the person ality of any particula r child but the child -
like state as such that is involved. Everything shows , therefore , tha t the ere
state is one of deperso nalization , whose role in the initiation process is
easil y understood. For thi s process consists, in practice, as much in losing
one 's for mer personality as in acqu irin g a new one, and th is loss is indeed a
necessary condition for this acq uisition process.
I would suggest, therefore , that we label this ere state a "tr ance (or state )
of dispossession. " And since the state is so closel y assoc iated with initi-
ation, we could further qualify it, among other things, as "a trance (or
state ) of initiator y dispossession. " One obse rvation made by Bast ide co n-
firms th at this qualifica tion is justified: with the impre ssio n made on him
by the no vices in the ere sta te in th eir "initiator y chambe r" stil l fresh in his
mind, he writes (1 958, 200 ) that "thi s childlike behavio r during religious
initiation" could well be "a survival in deepest Brazil of the childlike be-
Trance and Possession 49
havior observed in tribal initiation." Later on we shall have the opportuni-
ty to compare the behavior of novices in the Dahome y vodun cult with the
one observed in another region of Africa among a group of boys undergo-
ing a :r ib al initiation entirely unconnected w ith possession.
So what are the relations of this initiatory candomble trance to music?
The extremely detailed descriptions given by Gisele Binon-Cossard, who
never fails to record carefull y the part played by music in the ritual , an -
swers the question quite clearly: when the no vices are in a state of initia-
tory trance (ere state ), it so happens th at they occasionally sing, and when
the y do, they sing alone, unaccompanied by any in strument . When, on
the contrary, the y are in a possession trance (santo state ) they are always
totally silent, 105 and if there is music (whether vocal or instrumental ), it is
always made on their behalf by others, either by the men or women who
conduct the initiation, or by a group of drummers. The opposition of the
two states with regard to their relation to music is, therefore, abso lutel y
clearcut. I shall return on several occasions in later ch apters ro the music
specifically associated with possession tr ance in this initiation process,
and shall therefore limit m yself here to the music involved in initiatory
trance .
Twice a da y, morning and evening, the novices sing "the ingolosi pra yer,
which is very long and must be known absolutely perfectl y by the end of
the period of reclusion " (Cossard 1970, 168, 169 ). The author says no
more about this prayer in her book, but I know from her first-hand ac-
count that it consists in a ver y slow chant presenting specific musical char-
acteristics totally different from those in the remaining repertoire of the
candomble. Ever ything suggests that this sung pra yer is equivalent to the
long th anksgi ving song sung morning and evening by novices in the vodun
cult in Benin, which we shall discuss a little later in this chapter. Other rit-
ual songs, all much briefer, are sung by the novices during the course of
the da y : first upon awakening, then while going single-file to and return-
ing from the place of their morning ablutions, 106 before their meal, and
lastly at nightfall, when the y ask for lights robe lit (ibid. , 167-70 ). Since I
do not know the wording of these songs, I cannot say whether or not they
have any direct relationship with the initiation process, but it seems prob-
able since the y are in fact sung while in a state of initiator y trance. We can
be fairl y certain, however, that the y ha ve no connection whatsoever with
possession, since the y are never sung when the novice is in this sta te , or for
the purpose of entering it, or inducing it in his or her companions. This is
w hat we need ro remember for the moment.
We have a great deal less information on what happens during the initi-
at ion process into the orisha cult as practiced among the Yoruba in ige-
ria and, above all, in Benin. In his article entitled "The Role of Hebetude
[hebe tude] during Initiation to the Orisha and Vodun Cults," Pierre
 50 CHAPTER ONE

Verger does tell us, howe ver , th at "during the period between the day of
resurrection and the one on wh ich the no vice receives a new name [in oth-
er words during the initiation period], the latter appears to have lost his ·
reason and is plunged into a state of hebetude and mental torpor. . ..
While in this stare, the no vice is referred to as QmQtun or newborn
child. 10 ' It is in this state, when his mind has been washed clean of all pre -
vious memory, that the orisha's particular rh ythms, his songs, dances, and
all the god's forms of behavior, will be inculcated in him" ( 1954, 337 ).
This "stare of hebetude and mental torpor," with which we shall deal at
greater length later on, is clearly nothing other than the ere stare, as every-
one will agree. 108 Here again, therefore, I shall say that the stare involved
is one of dispossession. Verger does nor elucidate whether during initi-
ation the no vice alternates, as in Brazil, between the possession state and
the dispossession or initiator y trance, bur there is every reason to suppose
that this is the case . Else where, describing the ann ual ceremony for
Shango 109 as celebrated in Benin, Verger (1959, 52-53 ) records that the
man charged with embodying Shango throughout the entire ritual, which
lasts seventeen days, is po ssessed by the divinity only episodically. During
the remainder of the rime he is possessed, Verger tells us, by "a secondary
sp irit who accompanies the or isha" and causes the man to beha ve "like a
child or as if he were simple-minded, so that he smiles idiotically all the
rime." When in this stare he is treated wi th a familiarity very different
from the respect accorded him when he is possessed. In this childlike state
he is able to ear, drink, sleep, and fulfil his natural functions, bur "when
possessed by the orisha himself, his body must be totally exempt from
such prosaic obligations, " Verger adds. If this alternation of possession
and dispossession trance is observed during this comparatively brief rit-
ual, there is even greater reaso n to suppose it must also occur during the
initiation period which, to my knowledge is much longer. As far as this
particular aspect of the orisha cult is concerned , ~therefore, the process is
by and large the same in Africa as in Brazil.
Is the same true of music and of its relation s to rhe two forms of trance ,
possessional and initiator y or else possessional and dispossessional? It is
extremely probable, bur we lack sufficient information to be able to say so
categorically. We can, howe ver, say so with certainty in the case of the
very similar vodun cult. Therefore, we shall now turn to it.

Hebetude, Initi atory Dispossession, and Music in the

Vodun Cult in Benin
In the vodun cult of Benin 110 the length of the initi ation period varies from
one vodun to another, bur it is in general remarkably long. Herskovits
(1 9 3 8, 2: 181 ) mentions a tr aditio n according to w hich reclusion in olden

L
T ranee and Possession 51
times lasted eight years. In Porto -Novo, I was told about o ne vodun,
called Bodo, for whom the initiation lasted seven years . The lengt h an d
strictness of the requisite reclusion perio d discouraged the followers of
this vodun, however, so that to this day he has not made any new recruits
for the last forty years. Only ten years ago in the Porto -Novo reg ion, it
sometimes happened that the initiation period lasted for two years, and it
frequently lasted six months o r more . 111 The size of the new graduati ng
class - if one ma y use that term to denote a group of novices undergoing
initiation together - also varies a great deal. It can happen that onl y one
indi vidual is initiated at a time. Most frequentl y, howe ver, the new class
comprises between three or four and a dozen no vices. In 1969 I personally
attended the coming-out ceremonies of a graduating class consisting of
sixt y or so young initiates whose age varied from about two to about four -
teen years old. It was the largest I ever saw m yself , and it is worth descri b-
ing one ep isode of the ritual in detail for several reasons.
The vo dun involved was Ohwe , whose "convent" 11 2 is about two miles
outside Porto-No vo. When the first "coming -out " ceremon y had been per-
formed, the "new ones" 11 3 were convened , on an appointed day, to go to
pa y homage to the king. Led by their mon itress and escorted by women
carr ying their baggage , the y formed a long procession stretching for over
a hundred yards. Dressed in their finest clothes, heads newl y shaved,
wearing pearl headdresses, shoulders oiled, sh iny looking as if sculpted by
freshly inc ised scarifications, torsos barred by two broad belts of shells,
ankles encircled with metal rings, the y! 1• mo ved rapidl y, single -file along
the path, looking neither to the right nor the left. From time to time they
sang in perfect unison their walking song, a beautiful tune, ample, supple,
alert , and at the same time impregnated with a certain melanchol y. Hav -
ing reached the palace , where the king, seated on his cushioned throne,
aw aited them in the court of justice , the y prostrated themsel ves before
him and began to sing , presumabl y! 11 for the last time in their lives , a long
song o f thanksgi ving diligentl y learned during the ir reclusi o n. These no v-
ices m ade a superb sight , lying prostrate in a va st, spect ator -p acked inner
co urt , pres sed so tightl y against one another th at one co uld scarcel y make
them out save for their pe arl cro w ns and strings of purple-d yed cowrie
shell s undul atin g t0 the slow rh ythm of the melod y. The y sang the song
tw ice , w hich took about tw ent y minute s, after w hich the king addre ssed a
few wo rds of congratulation t0 the "new ones ." Tht mo nitre ss then tr ans -
lated his speech into their "secret langu age ," since the no vices, w ho had
not yet returned t0 ever yda y life, could neither understand nor speak an y
other. In answer , and still in perfect unison , the y sang a second song of
thanksgi ving , much shorter th an the first but in the same style . After that,
still at the order of their monitress , th ey withdre w inro another court of
the p alace , where the y were ser ved a large meal o f beans.
52 CHAPTER ONE

Several da ys later the very same ceremony was repeated in the palace by
the novices of another vodun, Siligbo, 116 who had likewise come to salute
the king. This time there were only four of them, but events unfol ded in
exactly the same manner. Indeed, this is how they had unfolded, or at least
h ad had to unfold not many yea rs ago, for all vodun in Porro -Novo.
This seemingly anecdotal aspect of initiation perfectl y illustrates the
problem at hand: In w hat state exactl y are these novices who sing on their
way to salute their king, and, once there , sing a very particular song, then
another, and then return to their convent ~till singing? The y certainly are
not in their normal state . Their robotic behavior, their totally "a bsent"
gaze, the fact that the y can neither underst and nor speak everyday lan-
guage, makes this quite clear. The y certainl y are not possessed, either,
since this state manifests itself by quite a different form of behavior. We
are thus forced ro assume the y are in the same state as the one exper i-
enced, as we have seen, by novices initiated into the orisha cult, both in
Brazil and in Africa; in other wo rds, here aga in we are de aling with initia-
tory trance. Howe ver, in the vodun cult this state displays particular
forms that we need to examine a little more closely .
We have just seen how no vices behave at the end of their initiation.
Now let us observe them at the very beginning of that process, as de -
scribed early in the twe ntieth century by Le Herisse in his historic work on
the former kindgom of Dahome y (1911, 135 ):
"Th e 'feticheur' who was showing us around his establishment
summoned his pupils. Approaching us we beheld three groa ning
creatures dripping with palm o il, crouching very low, one arm thrown
across their backs , the other pressed to their chests , fists clenched and
bent down at the wrists. Tw o of them were bedecked with flashy objects
and clothed in a simple, short, brightl y colored skirt; the third, more
recentl y admitted, had no covering other th an a piece of greasy cloth
around his loins. The y all kneeled down and, in shrill vo ices, recited a
monotonous chant that we were told was a greeting and a pr ayer for our
benefit. When the y had finished they lefr us, still groaning and
crouching, to sit in their hut, where the y imm ediately took on the most
bestially absorbed air we have ever seen .

Despite the quite colonial contempt inspiring this passage (wh ich causes
the author totally to misinterpret what he is seeing ), this description could
very well apply to no vices undergoing initiation as it can still be seen today
(if one is permitted to do so ), or at any rate as it could still be seen only a
few years ago. The photograph in Verger's article (1954, fig. 2 ) on hebe -
tude in the orisha and vodun cults provides a perfect illustration of this, '''
and the word hebetude is indeed the one that seems best to characterize
this particular state.
In the earl y stage of initiation-weeks or months depending on the
cases - this dazed state is apparently very pronounced, actually verging on
T ranee and Possession 53
total hebetude. The ne w recruits live completely apart from the rest of the
world in a "conve nt. " In theor y, the y are seen by no one other than the
priests and priestesses looking after them. Dressed in an unwashed, raff ia
loincloth , hair neither shaved nor dressed, the y are "dirty things," as I
have been told twice when I was permitted to see them in two different
convents in the Alada region. Heads stubbornly bowed, gaze obstinately
absent, they displa y absolute docility to their monitress. According to
Verger (ibid. , 33 8 ), this dazed scare is due, at lease partially, to the effect of
herbs and medicinal leaves whose use "is the most secret part of the rit -
ual." Let us say that the y, like the candomble novices, are drugged during
the entire initiation. 118 But as time passes, this giddy state becomes modi-
fied. From all appearances, the "new ones" I was able to see in certain con -
vents, who stood up straight and were no longer doubled over, had shaven
heads and were wea rin g extremely clean loincloths, were in a less deep
state of hebetude , repre senting, in all likelihood, a more advanced stage of
initiation.
In the vodun cult, initiator y reclusion is characterized by an important
ceremony during which the novices receive the scarifications that consti-
tute the mark of the vodu n or the vodun group to which they are being
consecrated. These scarificat ions , which are certainly symbolic but also
highly decorative, are incised on the shoulde rs, the neck, an d a section of
the torso. Blisters are produced by rubbing in certain powders, and the re-
sulting scars remain raised . By remodeling rhe bod y in this way, creating
the skin as sculptural material, profoundly altering the individual's ph ys-
ical appearance, "sratufying" it as it were, these scarifications are p art of
an even broader cosmetic undertaking which will acquire its full meaning
during rhe ceremonies char accompan y the novices' coming our from the
con vene, when they file our and dance in public, glistenin g with oil and be-
decked with their finest ornaments . As distinctive signs of rhe vodun,
however - and therefore in most cases of a lineage - the se scarifi cation s
can also be likened to bod y painting , scars, or tattoos char are done else-
whe re in Africa during initiation rituals and th at are destined ro display
member ship to a particular rr1be or clan, and ha ve nothing to do with pos -
session. A blood y and even dangerous procedure , these scarifications are
done in the secrec y of the convent, far from profane eyes. The novices are
drugged, I am told, in o rder to w ithst an d the ordeal, which rakes place in
an atmosphere of great excitement maint ained by continuous loud drum -
ming. An yo ne who has seen "fetishists" in D ahomey cannot have failed to
notice a particular kind of reserve, of dignit y char radia tes from them , an
inner bearing, as it were, that is different from th at of ordinary people, a
quiet energy that emanates from chem and distinguishes them from every-
one else. The long initiation the y had to undergo must certainly have pro-
duced a very special ripening of their personalit y. In this slow restructur -
ing of consciousness and character achie ved by vodun initiation, what is
54 CHAPTER ONE

the real importance of the scarification stage? We lack the data to answer
that question. But there is every reason to suppose that the novices emerge
from this orde al at least partly transformed. At all events, it is certain that
as initiation proceeds, the profound hebetude which they displayed at first
gradually gives way to a different state.
Once the initiation is concluded, the coming out of the "new ones" is the
occasion for a number of ceremonies, the details of which naturally va ry
from vo dun to vo dun. '' 9 Ar this point !er me simply say that the y are
sumptuous feasts all of \.vhich , in their different ways, are ostensibly in-
tended as a triumphal procl amation of the social group 's vitality. E very-
thing unfolds as if rhe intention were to show that the coming our of the
initi ates is a kind of victory for the group . Decked in finery and jewels, the
new initiates look like precious beings whose molding - one is tempted to
sa y gestation-accomplished ever so slowly and secretly in the heart of the
convent and now brought to term , is a dazzling triumph for which every-
one congratulates himself. Unquestionabl y, they no longer are "dirt y
things." On the contrary, the new initiates are now displayed as objects for
general admiration. Their yo uth , their beaut y, their wealth, their skill as
dancers, their capacit y to accomplish extraordinary feats - a subject to
which we shall return - are there on show to excite pride in all the onlook-
ers . True , their gaze is still abstracted and their docility still absolute; they
move like autom ata and beha ve like beings who ha ve become totally rei-
fied in the hands of those who have molded them. But their state is no
longer one of hebetude. The sta re of hypnosis in which the y seem to exist
appears, on the contrar y, to be a mark of their invulnerability, their indif -
ference to the external world, a sign that the y no longer have anything to
fear from it. (The childlike and totall y stereot yped poses the y adop t in cer-
tain circumstances, and described by Verger, do not alter thi s in any way :
un awa reness and innocence are manifestl y their lot. )
After their coming out, the "new ones" then spend several weeks leading
a life intermediate between reclusi o n and norm a l life. Every five or seven
da ys, depending on the calendar of their particular vodun, the y must re-
turn to sleep in the convent and remain there throughout the following
day, "·hich is the "vodun's pra yer day." T he rest of the time the y wander
about the co untr ys ide, visit ing houses and m arker s, performing their
dances, begging for food, and singing brief thanksgiving songs in re-
turn. ' 20 This is the stage w hen it is easiest to observe rhem. The y only go
somewhere under the guidance of the ir monirress, who is on the regul ar
staff of th e con vent and is also the one w ho looked after them during their
initiati o n. The y speak to no one else, see no one else , and obey her orders,
given in the cult's "secret langu age," like robots. The y give the impression
of being somewhe re else, of moving in another wo rld , bur in no way of be-
ing in a state of diminished or dimmed consciousness. On the contrary, as
T ranee and Possession 55
one watches them passing by single-file, peregrinating from one village to
another, what emanates from them is a sense of determination, compara -
ble co some extent co that of marching sold ier s whose eyes are fixed
straight ahead with never a glance co eithe r left or right. When the y per-
form their begging and thanksgiving dances and songs, wh ich they always
execute co perfection, they give the impression the y are somehow floating,
in a state of somnambulistic ra vishment. In fact they give the appearance
of being in a trance, but a restrained, internalized, reticent kind of trance,
one might say, a state ve ry different from the unbridled and paroxysmal
trances one usually sees during possession ceremo nie s. The y are clearly
not in a state of possession, but one is tempted co say that they are in a
state of grace. Seeing them, one thinks of those little Balinese dancing girls
described by Jane Belo (1960, 4 ), who perform in a state that she coo de-
scribes as somnambulistic. In this state, identified by a doccor as "hypnot -
ic trance" (ibid. ), these little dancing girls execute acrobatic dances the y
would be inc apa ble of doing, the Balinese think, "if they were not pos-
sessed by a god." Some of them, Jane Belo adds, sometimes "would dance
through hot coals without being burned." Yes, one may say, but these little
Balinese girls are in a state of possession, whereas the "new ones" I have
been describing are not. The moment has come, in fact, co consider their
behavior during the "co ming-out" ceremonies, when, ha ving recei ved
their new names, they are possessed.
Although, in this possession state, they do perform actions different
from those executed when they are in initiacory trance, the fact is that
their general behavior ne vertheless remains the same. In the case of certain
vo dun these ceremonies include a spectacular rite that is a kind of ordeal
by fire. 1 21 Outside, large ja rs-one for each initiate-containing food
floating in palm oil mixed with some other liquid are prepared for cooking
ov er as many separate hearths. At a given point the vapors produced by
the cooking process spont aneousl y ignite above the jar , an event that is
seen as a "miracle" (Herskovits 1928, 2 :165 ). Each new adept then goes
to his or her jar, thrusts his or her hand inside it , and removes the cooking
food without burning him- or herself, a sign that "the god is present. "
During this entire ordeal the behavior of the new initi ates does not differ
in any apparent way from that they display the next da y, or the follow ing
da ys, when the y wander from village co village singing their thanksgiving
songs. During these coming -o ut ceremonies they also perform dances -
usuall y in pairs , accompanied by drums, and some time s song - which al -
though not acrobati c are of sufficient virtuosit y co be one of the highlights
of the festivities. These dances form part of the repertory of possession
dances they will perform henceforth on various occasions, in particular at
their vodu n's annual feast. While executing these dances, during which
they are possessed, their beh avior is identical to that displa yed during the
56 CHAPTERO l

begging dances alread y described, performed when they are not pos
sessed: the y manifest the same abstracted air, the same robotic gait, th(
same hypnotized appearance. It is only some years later, whe n the y an
fully adult both as individuals and as initiates, that they give themselvei
up to the violent and extro vert dances in wh ich possession manifests irse!J
so plainl y, since it is expressed without restraint and can be recognized a1
what it is without a moment's hesitation.
If we thus consider initiation to the vodu n cults by obeying the rule we
imposed upon ourselves in order to study possession in general - that is,
looking at it as a dyna~ic process - we notice that its acco mpan ying start
of initiator y dispossession is modified considerably as initiation unfolds.
Between the stare of deep hebetude into which the new recruits are
plunged during the first weeks of their reclusion and the state of somnam-
bulisric grace they attain at the time of their coming -our ceremonies, a
great change rakes place. And indeed, w hat could be more natural? An up -
heaval of the self as total as the one the y experience during initiation could
hardl y take plac e wirhour a similarly coral modification of the altered state
in 'w hich the y live. Thi s altered state nevertheles s rem ains, from start to
fini sh , one of deperson alizat ion or, as I ha ve termed it , dispossession. Fur-
ther proof is that these "ne w ones" are given a nam e only after the y come
out from their reclusion and are solemnly presented to the outside world.
Thi s being so, are w e then reall y justified in say ing that this dispossession
is simultaneously a trance?
If trance is robe defined as an altered, rransiror y state of consciousness
conforming to a cultur al model, everything pro ves, as we have seen , th at
we are justified. True, it has the peculiarit y of being extremel y durable,
since it can last for several months, even year s. And the fact that it can last
so long is expl ained in part at least by the use of drugs. The fact remains,
ne vertheless , th at it is transitor y. Ther e is therefore no incompatibilit y
with the definit ion just propo sed. Even though some trances are very
brief, even verging on the moment ary, thi s does not me a n that others can -
not be very lo ng and follow, more over, certain spec ific stages of de velop-
ment. What is important ultimatel y is to make a clear distinction between
disp oss ession and possession tr ance. "A less violent trance " is how Cos-
sard describes the ere stare. True , but it is neither "sem itrance " nor "semi-
pos session trance, " contrar y to w hat Basride and Hersko vits have written.
What w e are dealing w ith is initiator y tr ance , and this should be under-
scored. In the candomble, as in vodun cults , initiation has two ultim ate
objectives, w hich are complementar y bur nevertheless distinct. The first is
to tr ain the initi ates, w hich means transforming them not onl y in terms of
their intern al structure (by conferring a cert ain knowledge and a certain
po w er upon them ) bur also in terms of their social relations (by m ak ing
them member s of a brotherhood w hich , in thi s case, can be likened to a se-
Tranee and Possession 57
cret society, since initiation is closely linked with secrecy ). The second ob -
jective is to create receptacles for the gods, by which I mean individuals ca -
pable of becoming possessed. Once initiation is accomplished, the initiate
will never be the person he or she used to be, and will never again bear the
same name, at least in Benin. But thi s does not mean that he or she will
permanently become the divinity to be embodied. The initiate will only be
possessed by that di vinit y from time co time. And even if he or she is never
co be possessed again, he or she would still remain an initiate. Even
though the goal of initiation is possession, training an initiate is, ne verthe-
less, an operation in its own right , and the initiate in his norm al state is
very different from the person he becomes when possessed. Initiatory
trance affects initiates, not those who are possessed . The fact that it is
easier, so it seems, co trigger possession tr ance in a subject already in a
state of initiator y trance (which is co say, in a state of dispossession ) than
in a subject in a normal state doe s not in any way suggest that initiarory
tr ance is an intermediar y state between the normal and the possession
st ate. :tis another form of tr ance, which sho uld be con sidered in and of it-
self and quite independentl y from po ssession.
Bastide, as we noted earlier, had al read y seen th at a connection could be
made between the ere state and "tribal initiation." Describing the gradu-
at ion of a group of initi ates among the Kissi, in Guine a, A. Schaeffner
( 19 51, 31 ) writes: "heads slightly lowered, these au coma ta seem to see
nothing aro und them. " Ever ything in his description reminds one of the
coming -out ceremony for the "new on es" of a vodun in Benin, with the
single exception that among the Kissi there is never for a moment any
ques tion of possession. Would the se yo ung Kissi bo ys also be in a state of
initiatory tr ance ? And could we not ask the same question in other cases?
The evidence leads us to believe we could. If this hypothesis could be veri-
fied, then initiarory tr ance might pos sibly appear as having no connection
wit h possession, an d as constituting a pa rticul ar and auronomous form of
tr ance , w hich is usually linked wit h religion , of course, but w hich , a
priori, could as we ll n ot be.
H oweve r different initi aro ry tr an ce and possession trance may be from
one anothe r, they nonetheles s constitute, at least in vodun cults, a contin -
uum between whose two extremities all the intermediate states obs er ved
occur. I must stress, however, that this does not me an that the distinction I
ha·1e tried co establish between th em is either none x istent o r even purel y
convenie nt . It is just as real as the one that differenti ates a "u" soun d an d
an "i" sound, a distinction th at in no way pre vents us from moving from
th e first ro the second, and vice vers a, without being able to say exactly
whe re o ne begins and the other ends. The phenomen a we are studying
here never display oppositions of the type th at T roubetzko y, in the field of
phonology, termed "priva tive ," whi ch is co say th at the y are based on the
58 CHAPTER ONE

presence or absence of some distincti ve feature or, if one prefers, on an

"all-or-nothing" basis. The oppositions we are dealing with here are .
"gradual" ones, in other words the y operate on the basis of a "more-or -
less" type of distinction and this is what frequentl y makes analysis so diffi -
cult and uncert ain. This me ans that an y valid interpretation of the facts
can onl y be m ade from a structural viewpoint, that is, by taking into ac-
count the whole situation in wh ich the phenomenon occurs and the system
w ithin which it is at work. During initiation, the novice w ill probabl y have
mo ved very graduall y from the profoundl y gidd y state into which he or
she sinks at the beginning to the state of inspired som nambulism displa yed
at the end . The same hold s true of the following phases. The transition
from initiatory disposse ssion tra nce to posse ssion tr ance experienced dur-
ing the ordeal by fire described earlier, for example, also seems to be very
gradual; in an y event, it is imperceptible to the spectato r. During a period
as ch angea ble as the end of initiation, it is also very difficult to say which
particul ar state is involved at an y point. Moreo ver, it often happens in
Benin that the mo st eviden tly genuine pos session trances manifest such
discreet external signs th at they go totall y unn ot iced by the uninitiated . It
is not surprising, therefore , that, in the c..asew ith which we are dealing
here , external beha vior is insufficient co identi fy the type of tr ance in-
vo lved .
H aving stated these p oints, w e can now ask o ur selves in w hich way ini-
tiator y (or dispo ssession ) trance and possession tr ance differ in their rel a-
tton to music.

T he Musical Repert oire of Initi ation

During the entire initiati on per iod a large part of the no vice's time is de -
vo ted to dail y musical activit ies. These can be divided into two quite dis -
tinct repertoires. On the one ha nd , there are a cappella songs th at the no v-
ices sing themselves, and th at no one else ever sings; on the ot her , there is
drum music, so metimes accomp anied by singing and some times not ,
w hich is executed by specially app o inted musici ans so th at the no vices w ill
dance. The fir st repertoire is made up, for the mo st part , of chants clo sely
connected w ith the initi ation itself, and w hich the no vices will ne ver sing
aga in once their initi atio n is ove r an d the y have graduat ed from their nov-
ice st atu s. T his w ill be labeled the initiaror y repertoire. The second reper-
to ire, on the contrar y, comprises the rh ythm s and tunes that will be pl aye d
and sung at all the cult's public ceremonies , p articularl y during the vo -
dun 's annual feast , in order ro trigger and accompan y the possession
dance. This will be labeled the possession repertoire .
The sa me kind of distribution of the music can also be observed in the
candomble . On the possession side w e find , fir st , the invocations sung by
T ranee and Possession 59
the "saint's father" while he rings the sacred bell in order to trigger the
santo state, and second, the drums and songs used to incite the no vices to
dance ;,hile they are in the santo sta te. On the initiatory side, we find var-
ious songs executed by the novices in their ere state, and in particular the
long prayer, ingolosi, that we mentioned previousl y. This chanted prayer
is characterized, as I said earlier and w ith out being able to go into greater
derail, b y a particular musical style completely different from that of the
singing heard in the public ceremonies. This same distinction, between
initiator y and possession repertoire, can also be found, as we know, in the
vodun cults; but in the latter case, I wil l be able to describe their contents
in rather more detail.
The long thanksgiving song, sung when the "new ones" are prostrate in
front of the king, is ' " performed during the entire reclusion period by all
the novices twice a day, at dawn and just before sunset, except on the "day
of the vodun" or, in the case of Khevioso (god of lightning ), unless he has
thundered that day. Each vodun, or vo dun family at any rate , has his or its
own particular song or chant, bur all these so ngs display the same stylistic
features, which are extremel y m arked . 123 Wi th out going into the musico -
logical derails , let me say first of all that the long thanksgi ving chant (since
there is a short one too, which I shall come to later ) is performed using a
type of vocal emission totall y different from the ordinary singing voice
used in this region of Benin. Thus when singing for the vodun Sakpata 12 •
(god of the earth ), one sings in a fluted voice and in a high -pitched register.
The song is purel y vocalized -me aning that no actual words are in -
vo lved - with very long note values (whole notes, half notes ) that are con-
nected by melismas using successions of short or very short notes (quarter
to sixteenth notes ). All notes ma y be dotted or nor, but the melismas are
always sung - one would like to say drawn-with the greatest precision.
The voice completely lacks vibrato, which gives this very slow melody,
punctuated by carefull y measured silences the aspect of a zen painting ,
whose spirals would va nish slowly into the air even as the y were formed.
The song, which lasts for about ten minutes, is made up of a sequence of
seven stanzas, each repeated two or three times with va riations. The word
"hieratic," had it not been so misused, would perhaps best characterize
this music, which is so impregnated with spirituality, so otherworldly, th at
it con veys an ove rwhelming sense of strangeness. Nothing could be less
like "black" music than this. Th e voices, so light, so purged of materialit y,
so disembodied, mo ve through space w ith out any rh ythmic support and
create a totally abstract architecture . The same melodic motifs are em -
ployed throughout, but linked into seq uences that are constantl y va rying,
so that one never quire knows where one is. Ir is an immaterial, seraphic
music , suspended bet w een heaven and earth, as far removed as it can be
from the music one usu ally hears in chis region, a sound that creates a feel-
 60 CHAPTERO E

ing of complete remoteness and of projection into another world. Yer it

conveys at the same rime a sense of great perfection, both in form and in
execution. The perfection of art. If the phrase were not pejorative, one
would be tempted to say the perfection of artificiality, since nothing could
be further removed from nature than this singing . Ir begins with a long,
held note sung in the very highest register of the voice, 125 and almost cried
out, it then slides very slowly down to the semitone below. The chromatic
effect is striking. It ends on the same note, sung to conclude the repetition
of the first stanza in a mirror form da capo, which completes the extraor-
dinary impres sion of aremp oral ity chis music produces. Everything is done
to conjure up a sense of the unusual, to create by musical means a world
apart, a world far removed from chis one, as far apart as the life the nov-
ices lead during their reclusion.
The strangeness of chis singing, its formal perfection, its impassibility,
the disembodied timbre of rhe voices, the atemporali ty of the melodic de-
velopment , the total abse nce of meaningful words -all chis gives the music
an abstract, almost geometric character. What does this mean? Obviousl y,
nothing could be less gratuitous than this sort of Greek temple, as it were,
of African religious song. For me - and here we inevitably enter the realm
of total subjectivity-this music is the very expression of the zero state of
the self, of depersonalization, in short of dispossession. This is the music
of a being somehow floating between essence and existence, and it suc-
ceeds admirably in its purpose, which is co express the experience of dis-
possession.
Apart from this long thanksgiving song (which is always sung twice in
successi on, so char the performance lasts more than twenty minutes and
constitutes an amazing vocal feat, since the voices are under extreme pres -
sure from beginning to end ), there is also a much sho rter one w hich is sung
at ocher rimes and in ocher circumstances. In chis case the song has words.
Its melodic structure - in accordance with the brevity of the piece, which
lasts anywhere between twenty and forty seconds depending on the case-
is totally different from char of the long one; bur it too makes use both of a
certain chromatism and a, let us say, bizarre method of voice production,
two features chat conjointl y produce the same feeling of strangeness. 126
Different as the y might be. from one anothe r, these two thanksgiving
songs, the long and the short, therefore unquestionably belong co the
same repertoire. And the same is true of the third category of so ng , al-
ready mentioned, which is sung like the shore thanksgiving songs during
the period when the young novices have emerged from their reclusion and
proceed from place co place begging for food. Unlike the thanksgiving
songs, however, wh ich are sung in a prostrate position, the begging songs
are executed while dancing, although, like the former, they also make use

L
Trance and Possession 61
of chromatism, which, I must repeat, is extremely unusual in the music of
this reg ion.
Both the thanksgiving and the begging songs, which constitute the es-
sential core of the in itiator y musical repertoi re in vodun cult, are thu s
characterized by musical features that differentiate them from the music
of the Fon and the Gun as a whole. T he beating of the drums and the songs
that make up what I have labeled the possess ion repertoire belong, on the
contrary, to the region 's habitual musical language. In a later chapter we
shall see that this is a general rule, and I shall go inn the reasons for this. If
these two repertoires differ so greatly, both in their musical structure and
in their function (the initiatory repertoire is sung by the initiates whereas
the p ossess ion one is pla yed for them ), this is because the y are concerned
with two distinct aspects of the cult, initiation on the one hand, possess ion
on the other. The latter, possession, is directed toward the outside since it
has no meaning unless it is public. Initiation, on the contrary, is directed
inwards. It is the secret side of the cult. Music of secrec y, this initiatory
repertoire is kindred to that music of masks and of secret societ ies which
occupies a very distinct place in African music as a whole. The bizarre
vo ice production and the chromarism by which they are characterized
seem to be somewhat similar to a vocal mask and a disguise of the person .
In the case of possession, on the contrary, it is essential to know by whom
the subject is possessed.
The role pla yed by initiatory music in the vodun cult posed a problem
made even more interesting by the fact that this cult occupies an important
place in the general picture of possession cults. We therefore could ask
ourselves if the unusual musical devices it brings int o pl ay- its strange -
ness, in short-could be directly connected to that equally strange event
we call posse ssion; if the disorienting or distancing effect it produces could
be cultivated in order to trigger char separation from the self, that loss of
the self in the beyond experienced by the possessed person. We find, in
fac t , that this is not the case at all. In practice, this mu sic has no connec -
tion w ith possession, or at any rate only indirectl y. It is never associated
with the triggering of possession trance, which it never pre cedes and with
which it is never sim ult aneous. This being so, one might say it would have
been enough to state it, certainly. Bur we also needed to understand the
wh y and the whe refore of the matter. Why this music then) Because the
vo dun cult is not solel y a possession cult. It is just as much an initiatory re-
ligion and a social institution that lays great stress on the integration of age
groups. Thi s music is not assoc iated with ap prenticeship co possession; it
is assoc iated with the tr ain ing of the initiate. But the two are linked . Hence
the importance of the dispossession trance wh ich , in the context of these
very long init iations, constitutes a sort of backdrop that makes the airer -
62 CHAPTER ONE

1
nation of initiatory and possessional practices possible. " The function of
this initiator y music is both tO express this dispossession and to assist in .
bringing it about, since in and of itself it manifestl y constitutes an ascesis.
Hence its strangeness. As for its beaut y, I shall not vent ure to attempt an
explanation. It is a kind of beaut y that the art of music can achieve w hen-
ever it is at the service of religious fer vor , and w hen that fervor represents
for the w hole of a societ y the most essential form of vita l force.
2 Music and Possession

THE WORD "M USIC"

"Ever ything is music," John Cage once said. If this were true, this book
would have been pointless from the scare, since the relations it seeks to
bring out between trance and music would necessaril y be relat io ns to ev-
er ything, or co anything. Bue this is in fact untrue, o r rather it is neither
true nor false. There are some statements char should nor be taken more
seriousl y than thei r authors expected. Everything, silence and emptiness
included, can in effect be heard, or seen, or felt, as a piece of music m igh t
be : iris ultimatel y a matter of good will. A squeaking door, a faul ty loud -
speaker, the sound of a flushing toiler can a ll be appraised with a musical
ear . Some find it pleasurable and if the y sa y so we muse believe them .
Equ ally, nothing prevents one from saying chat a landscape can be viewed
as music for rhe eyes. If "ever ything is music," then we are dealing with
metaphor , which is a perfecrl y legitimate wa y of using wo rds and doubt -
lessly as old as rhe world; but I shall attempt to avoid it since iris of m ini -
mal profit for our enterprise.
What will the word "music" mean in chis book) Given its context, char
of trance, the word music will be used to sign ify any sonic event tha t is
linked with chis scare, char cannot be reduced to language - since we
would then ha ve to speak of words, not music - and char displays a cer-
tain degree of rh ythmic or melodic organiz ation. Music will therefore be
taken in its most empiric al and broadest sense. In ocher words, it will not
be treated here as an arc but as a practice displa ying the greatest possible
va riet y of aspect s , extending from the most discreet rustling sounds pro -
duced by rhe shaking of a basketr y racrle to rhe deafening unleashing of a
la rge group of drum s , from rhe solit ary tinkling of a tin y iron bell to rhe
o rchestral splendors of a Balinese gamelan, from th e most elementary
mon od y, chanted recto tono , to the most comp lex voc al pol yphonies
(such as chose of the Bushmen ), fro m the line arit y o f a simple motif pro -
duced by clapping hands co rhe extremel y comple x interpl ay of a large per -
cussion ensemble, from the refinement of a viol in rune skillfull y varied by
a professional virtuoso co the rustic sound of a summons bellowed
through an animal horn or capped out w ith a nail on rhe side of an empty
beer borrle.
'!
64 CHAP T ER TWO

ow rha r I have empi rica lly del imit ed the field covere d by the wo rd
"mu sic," we sh ou ld as k ou rselves whet her the concept at wo r k is or is n ot
ade qu ate to th e kin d of inves tiga ti on I am unde rt ak ing. In p r acti ce, such a
defi niti o n emer ges from the delim ita ti on of a conc rete wo rld of so und th at
is totally foreign to the o ne p rac t iced by the majo rit y of the soc ietie s in
w hi ch the fac t s we w ill co nside r can be obse rved. 1 By wo r king w ith a co n -
cept of this kind we are thus tu rn ing our backs on the system of th o ught
proper to these societ ies . Bur let me no r hedge o n this po int ; th is book is
no r aimed at t hese systems. or that I igno re nor w ish to m in imi ze the ir
import ance or their intere st; quite the cont rary. If, in the case of certai n so -
ciet ies at least, we possessed a t rul y indigenous taxono m y of not ions relat-
ing to what we te rm m us ic, ' and a matching theor y of how t hose n ot ion s
rela te to t he trance state , I wou ld cert ainly use it as the p rim ary bas is fo r
my resea rch. T o m y knowledge , howeve r, no such documentat ion ex ist s.
Since we are unable to conduct o ur anal ysis with indigenous co ncep ts, we
must therefore use our own . Ir must be no ted , howe ver, that this would ul-
ti m atel y ha ve been the case anywa y. A demon str ati on , had it been poss i-
ble , of the relat ion s a p articular system of thought - be it African, Asian ,
or European - establishes be tween trance and p articular sound events fall -
ing with in our definition of mu sic, cou ld ne ver hav e constituted more than
the first stage of our work . M y aim can onl y be to brin g out these re lat ion s
by go ing be yond interpretat ions necess arily belo nging to cogniti ve pro -
cesses different both from each ot her a nd fro m o ur o w n. Whate ver ma teri-
als are used, whate ver deg ree of elabor atio n is reac hed , wh ate ver langu age
is used to co nve y th e finding s , we will alv,:avs h:ive to reinterpret them in
our own terms. When I am to ld by a Son ghav musician in N iger , via the
wo rk o f Je an Rouc h, th at he tr ansmit s hi s mu sic to th e ear s o f the sp irits by
me ans of a rattling de vice extending fro m the neck o f his single -str inged
fiddle ,3 I rake note of this informati on , w hich is indeed of the greatest in -
.· :· terest ; but I must then go on to reint erpret it w ithin the framework of a
more gener al system o f symb olics . W hen I rea d , in the w ork of H enr y
Sko ff Torgue , th at for pop mu sicians t he doub le b ass "lend s securit y" w ith
its low notes , th at the percu ssion is "ne ver fren zied ," and that it is the
voice , o r th o se instrument s derivin g fro m it (in o the r wo rds, those of me -
dium or high pitch ), th at furnish the "hysteri ca l crie s," I am then led to
consider the se obser va tions , w hich are scatter ed through o ut his book, as
signif icant o f an underl ying system . W het he r it is cle arly expres sed , im -
plied , or ign o red altogether - which is usu ally the cas e- is o f lit tle import.
Just as I am w ithin my right s to include w ithin a single categ ory- the o ne
o f "mus ic" as I hav e just defined it- fact s th at other s wo uld ass ign to dif -
fer ent categorie s, so am I at libert y to o rgani ze into a system facts that oth -
er s would not link toget her at all, o r wo uld orga ni ze on the b asis of differ -
ent viewpoi nt s. T he question is to kn ow if the result of such a
rear r ange m ent is logic ally co nclusi ve o r nor. The re ader w ill be the judge.
Music and Possession 65
Any music can be considered from three different view points: first, in
and of itself, as an object (independent of its maker and of anyone listen -
ing to it ); second, as something produced either by composition or by ex-
ecution (as a subjecti ve creation ); third, as something listened to (from the
viewpoint of perce pti on ). To use the term s proposed by R. Ja ckobson
(1960 ) for the analysis of language , mu sic may be considered from the
viewpoint of th e "message" it constitutes, from th at of the "a ddresser"
who tran smits it, and from that of the "a ddressee" who recei ves it-or, if
one prefers, from that of the "transm itter " and the "receiver."' In the nor -
mal conditions of mus ical communication, the transmitter and receiver
share a common code , so that the message sent by the former refers to a
context understandable by the latter , and communication is established
between them by the action of a cert ain contact. "Each of these six faccors
(transmitter, me ssa ge , recei ver , code , context, contact ) gives rise to a dif -
ferent linguistic function," Jakobson tells us . Although we must guard
ourselves from systematically applying w hat has been said about language
co music, since the on e can by no means be reduced co the other, the same
proposition can nevertheless be made for music , and w ith equal chances
of leading to interestin g observat ions . Throughout thi s book we will see
that, depending on w hether the music is made b-:;•the person in trance or,
on the contrary, for that person , music and tr ance have quite different re-
lations co one an other.

WHEN D OES M USIC ACT'

We ha ve just seen that pos session should be viewed as a dynamic process .

Therefore, we will begin by asking the question, When does music act?
The individual in the process of experiencing tho se ph ysiological, psy-
chological, or emotional disruptions chat will orient him coward posses-
sion w ill, generall y speaking , gra vitate coward the milieu of possessed per-
sons , and will consequentl y be present at their ceremonies. Since these
ceremonies are accompanied by mu sic, he becomes impregnated with the
musical atmosphere of such possession sessions. But the chronic crisis he
is ex periencing and the acute crisis th at w ill follow it most of the time do
not necessaril y have any direct relation co that music. Indeed , the y often
have no relation co it at all. As Loui s M ars very rightl y points out (1953,
225 )- with specific reference to Haiti an voo doo ( but his observ ation
could as w ell be applied co other cult s)- "We have co distinguish the ritual
posses sion cr isis from the possession sickness that arises (generally speak-
ing ) independentl y of any ceremonial atmosphere ." Music can undoubt-
edly sensitize a sub ject to the call of possession, and can thus contribute co
awak ening his vocation; this, howe ver , does not mean th at music is re -
sponsible for the ps ychological disturb ances he encountered and whic h
led him co follow this path .
'E 66 CHAPTER TWO

In Bali, Jane Belo ( 1960, 48 ) describing the initiation of a medium, tells

us that it was as the result of a series of trances - in my terminology crises
or fits - undergone in va rious places and circumstances (but always out -
side of an y ceremony and thu s in the absence of music ) that the person
concerned, after having consulted a priest on the meaning of these at-
tacks, became convinced that "t he gods were calling him ." In this case we
are clearly in the presence of what we have labeled the prepossession cri -
sis . Similarly, J. H. Nketia (1957, 5 ), writing about possession in Ghana
(which he interprets as being essentially a renewal of m an's alliance w ith
the gods or with his ancesto rs), observes chat "individuals are known to
gee possessed in pri vate life outside the co nte xt of mu sic and dancing. " We
may suppose chat here again we are dealing with prepossession crises. In
any case, let us remember th at such crises can t ake place without music
and without dancing. Among the Dogon , future prie sts of the binu (myth-
ical ancesto rs) undergo crises or fits that occur unexpectedly, not only dur-
ing ceremonies that include music-and are not necessarily rel ated to binu
worship' - but also ou tside of any ceremon y and consequently with no
music at all. • This is no longer the case afte r their o rdination, however,
and I shall return to this point later (p . 98 ). In the candomble, on the other
hand (Cossard 1970 , 158 - 59 ), it is usuall y wh ile the future initiate - of
either sex -i s attending a ceremony chat he or she is suddenl y struck by
the divinit y and collapses to the ground. Thi s is a sign chat a god has cho-
sen him or her as his "horse. " But at chis entirel y preliminar y srage of pos-
session, neither the victim nor an yone else yet knows which god is in-
volved . Contrary to what happens to those who are already initiated, and
who go into trance only when summoned to do so by the song or rh ythm
specific to the god who is to inhabit them , here the person cannot have an -
swered such a summons since he or she does not yet know the code. Gisele
Cossard further points out, as we saw earlier, chat the "candidate" can
also be struck an d fall in this way at any time outside of any ceremonial
context. Thus , despite appe aran ces , prepossession crisis must be regarded
in the candomble as well as being independent of music .
In Chad, among the Mussey, ' it oft en is while she is out in the bush that
rhe person abou t to be possessed is seized by an unknown god and struck
ro the groun d. She will then lie where she has fallen, unconscious, until
women belonging to the college of the possessed come near her and blow
into a gourd trumpet until the right notes awak en her. In this case then
music does not trigger the crisis , but on the contrary, terminates it, by es-
tablishing communic at io n with the god who is responsible for it.
Among the Hausa , noniniriared girls experience possession crises by in-
gesting a drug, betel darura (J. Monfouga-Nicol as 1972, 200 ). This "pro -
fane possession" displa ys all the symptoms of normal possession crisis ex-
cept for the fact chat "the gods clearl y do nor manifest themselves." The se
82 CHAPTER TWO

larly in Tunisia, in the case of the possession dances performed by women

as they listen to the dhikr chanted in the neighboring room by the men, it is
when the rh ythm of this chanting accelerates th at they begin co dance and
fall into trance (Ferchiou 1972, 56 ). The same is true of other marabour-
type sessions in which the music, wh ich is quite different, is in fact true
possession music, since it is played specifically for the purpose of inducing
possession . Here again, the session has two phases: "during the first, the
at mosphere is relaxed, the rh ythms are slow, and the words distinct; this is
the phase in which all those who are present participate . The second be-
comes progressively more intense, and the rhythms, which steadily accel-
<'Tate, are those of the possession dance" (ibid. , 57 ).
In Tibet , the invocations chanted in order co trigger the o racle 's trance
and addressed co the divinity who is co descend into him, are first chanted
slowly, but soon "the tempo [is] increased" (Nebesky-Wojkowitz 1956 ,
433 ). In Bali, Jane Belo ( 1960, 35 ) describes how the orchestra accelerates
suddenly in order co trigger the trance of the barong carrier. One could go
on and on listing such examples. But even this rule, needless co say, is not
without exceptions. To mention only one, again in Bali, the orchestral
music accompanying the trance of the kriss dancers at the end of the bar -
ong ceremony maintains an even tempo from start co finish. 0 0
As is frequentl y the case with all sorts of music, acceleration of tempo
ver y often goes hand in hand with an intensification of sound. The record
entitled Fete pour l'offrande des premieres ignames aShango (Feast for the
offering of the first yams co Shango; G. Rouget 1965 ) provides us wit h a
typical example of simultaneous acceferando and crescendo; in this case
the acceferando is achieved not so much by an accelerat ion of the tempo as
by an incre ase in the number and insistence of the drumbeats within it ; as
for the crescendo, it results both from the intensification and dram atiza-
tion of the singing and of the horn call. In Bali, describing the ceremony
during w hich "the official entrance of the god" into the medium who is co
embody him will take place, Jane Belo ( 1960, 50 ) observes that in order to
encourage trances in the mediums present , people were sing ing loudly
while the orchestra played "as loud as it could." Elsewhere (ibid., 20 ) she
notes that, aga in in order to trigger the trance state, "the singing and the
music of the orchestra redouble their intensit y." Among the Thon ga of
Mozambique, H . .A..Junod (1913, 438 - 43 ), describing a session held for
the purpose of inducing the sick man co declare the name of the spi rit pos-
sessing him, emphasizes the violence of the percussion, which is composed
of single -skinned drums, drums made of large tin cases, and calabash rat-
tles shaken so as to almost "slightly couch the he ad or the ears of the unfor-
tunate sufferer," the total result being "a hideous din." In add ition to these
instruments there also are "incantations" remarkable for the singula rly in -
cisive an d penetrating nature of their melod y, whose disturbing effect is in-
Music and Possession 67
false possessions o ccur when girl s, who wa nt to imit at e bo ri adepts, swa l-
low d a tur a seeds in order to do so (ibid., 199 ). But formerly they a lso use d
to occur in a more institutionalized way . On particul ar evenings when the
girls congregated to sp in cocton together, they amu sed themsel ves by in -
troduc in g the yo un gest amo ng them to the into x icat in g effe cts of the betel
datura. "When the girls began to shake, dribble, a nd spin endlessly
around and around, mouths filled with froth, th e yo ung people's griot 8
came to pla y the plant's mocto on his d rum, then immediately aft er wa rds
that of the god Kure. 9 Th e girls we re th en told to sit down, given cool milk
to drink, and the griot tapped them on their heads w ith hi s drum. The cri-
sis subsided and o rder was re stored." In thi s example music is used not to
trigger the crisis bu t , on the contrary, to m ake it subs ide , w hich corre -
sponds tO what was just mentioned fo r the Musse y. The crisis subsides
owing tO the symbolic all iance effected with the god and ob ta ined, first,
by be a ti ng out his mocto, second, by bringing the drum, th e sac red instru-
ment, into contact w ith the head of the falsely possessed person. W hen it
is simply a case of fa lse possession in littl e gir ls imit ating the grownups'
bori, also induced by datura, the you ng people's griot is likewise sum-
moned. H e "plays his drum and touches the girl's he ad with it. She then
calms down and dances tO the sound of th e music. E ve ryone knows th at if
the gri ot w er e not summoned, th e child 's belly would swell up and she
would die .... " A H ausa woman ex plain s the situ at ion as follows:
"Whe n the y go inro tr ance yo u h ave to call fo r the gri ot and he pl ays his
drum; the g irl s dance like fools. It is because the y d anc e th at the babbajiji
(the dacu ra ) leaves them. " H ere again, we find that musi c is used to end
the crisis an d not ro pro voke it, but thi s time it does not act symbolically;
it spurs the girls on ro dance.
Exactly the same thing hap pen s in tarantis m , as we know from the fol -
lowing text dating from 1621 and quoted by E. D e M a rtino:
True, the action of music is miraculous, and if one had nor seen it with
one's own eyes, one "·ould never believe it possible th at the person who
had been bitten by the tarantula, almost moribund because of the
venom, moaning , angu ished , agon izing, almost bereft of his senses,
external and internal . . . , [could ] come back to his senses as soon as he
has heard the sound of musical instruments, open his eyes, prick up his
ears, rise, first begin to make slight movements witli his fingers and roes,
and then , keeping the rh ythm of the melod y, which he finds pleasant and
helpful to him, start dancing with the greatest liveliness, gesticulating
\\·irh his hands , his feet, his head, with all the parts of his body shaken in
all of its limbs by the most varied agitations . [1966, 385 )
In the Haus a bori, dance is used as a means of cou nter ac tin g the effects of
the betel datura; in ta r antism it cures the bite (re al or imagi n ary ) of the
spider. We should remember th at in both cases the function of the music
68 CHAPTER TWO

involved is not to trigger the crisis or fit but, in fact, to cure it, to provoke
its resolution in dance. All the same, there is one notable difference be -
tween these two cases. In the bori we are in fact dealing with a false pos-
session dance , which prefigures the real one by skipping the initiation
phase. In tarantism , in whic h the crisis leads directl y t0 possession with-
out skipp in g this phase (since it functions without initiation by nature ),
we are dealing, on the contrar y, with re al poss ession dance.
It should be obse rved in pas sing that this absence of initiation, which is
perhaps tarantism 's principal characteristic , is undoubtedl y due to the fact
that it is not a completely socially integrated religious practice. Since there
is no initi a tion, the prepossession crisis is never en tirel y re solved, w hich is
w hy it recurs peri odic ally. Thi s would ex plain the fact th at in rarantism,
contrary t0 the general rule , the prepossession crisis is not a unique event
in the p oss essed per so n 's life but an annual recurrence.
In the ndbp, prepo ssession cr isis, or, to use Andras Zempleni 's term
( 1966 , 314 ), "non -ritu alized poss ession crisis ," of ten occurs during a cere -
m ony, but can equ ally well occur w ith out a ny cerem o ny and thus in the
ab sence of music , 10 which conform s to everything we have said up to now.
On the other h and, as we have seen, the no rm al clim ax of ritu al ized pos-
session trance in the ndop is a crisis re so lved by a fall to the ground, a nd in
this respect it has the same exterior signs as preposse ssion crisis. This time
both crisis and coll apse are directl y related tO mu sic, as we know from
Zempleni's (ibid., 358 - 60 ) description and what he tells us about the be-
h avio r of a new ndbpkat ' ' at her first public seance. She mu st, "if possible,
achie ve a crisis leading tO collapse . If necessar y, she will be dragged in
front o f the drums in o rder t0 force this outcome" (ibid. , 37 5 ). So whe re as
the "uno rga nized crisis" (ibid. , 415 ) a nd the en suing collapse do nor seem
t0 be necessaril y related tO mu sic, the ritu alized crisis and collapse, on the
contrar y, appear tO be directly pro voked by it.
Let us now p ass on to initi at ion. The role of music seems tO differ ac -
cording tO whether one is dealing with the bori, for example, or the can -
domble ."
In the bori, possession is always acco mp anied by mu sic throughout the
entire initi at ion, both durin g the exorcism ph ase (d ark gods ) or the ensu-
ing imp os iti on phase (w hite gods ). 1 ' In the candombl e, on t he contrar y,
the m ajo rit y of posses sions that occur during initi ati o n do so without mu-
sic, or almost so. As we saw ea rlier , ' " the novices live in the ere state
th roughou t their reclusi o n period, and thi s st ate is interrupted by short
peri ods whe n the y go into the santo st are , which is tO say inro a possession
tra nce . The day begins w ith a mo rning purificati o n bath. Wa kened before
dawn, the y walk to their bathing place in silence w ith their "little mother "' 5
walki ng ahead of them rin ging the adz a ' 6 or sac red bel l. Each in turn is
bathed w ith cold wa ter by the "little mo ther. " The shock the y experience ,
Music and Possession 69
Cossard writes ( 1970, 168 ), is usually enough to ensure a return to the
santo state. There is every reason to believe that the bell is not rung d urin g
the bath ceremony itself, so that here we are faced with possession occ ur-
ring in the absence of music. There is another possession trance that oc -
curs without music: the one caused by drinking abo -an infusion of leaves
collected from plants that h av e the "property of pro vok ing and reinfo rc-
ing the trance state" (ibid., 181 ). As with borz, the candomble thus pro -
vides examples of trances induced solely by the action of drugs and in
which music pla ys no role whatsoever.
Continuing with the candomble, at least seven days after the beg inning
of the reclusion period, a first ceremony takes place, at the end of which
the no vice become s a jawo, or the orisha's wife. "The saint's father sum-
mons the orisha with the appropriate salutations while ringing the adza.
This is the dramatir moment of the ceremony, for it can happen that the
orisha is slow in coming" (ibid., 176 ). This possession, which marks an
import ant phase of the initiation process, is thus accompa nied musically
by nothing more than the tinkling of a sacred bell.
Several days later another ceremony takes place in which the jawo,
while in the santo state, is bathed in blood in order to establish the m ysti-
cal bond that will henceforth bind her to her orisha. The ritual, Cossa rd
tells us (ibid., 184 ), is identical to the assamento ceremon y she descr ibes
elsewhere, which means that it too takes place in the complete absence of
music. On the evening of the same day the naming feast rakes place in t he
presence of a large audience. When the jawo make their appear an ce they
are already possessed . Upon the invitat ion of a woman who accompanies
them and to whom the "sa int's father" has entrusted the adza, the new
jawo then cries out her name. At this point the drums are beaten "to greet
the birth of a new orisha" (ibid. , 185 ). Thi s is anothe r ex amp le of a pos -
session trance assoc iated wi th the sacred bell but owing absolu tel y noth -
ing to the sound of drums.
Among the Wolof, as among the Thonga, it is 1t the end of a ritual in
which music plays a central role that the possessed person is induced tout -
ter publicly the name of the divinity possessing him. This is, of course, a
crucial episode in his initiation, since by uttering this name, the possessed
person not only identifies the divinity but, in so doing, enables himself to
become identified with it. Ther e are broad analogies between the kind of
music used in the two cases . Zempleni (1966, 356 - 60 ) in his work on the
Wolof and H . A. Junod (1913, 438 - 43 ) in his study on the Thonga , both
place equal emphasis on the viole nce of this music, the purpose of wh ich is
to reduce the sick man to a state of nervous exhaustion: "almost intoler-
able noise of the drums," Zempleni w rite s; "hideous din," notes Junod.
We shall later examine what exactly this music consists of. For the mo-
ment we need only recall that in both cases music attempts to induce the
70 CHAPTER TWO

crisis through the same method. "Conquered at length by this noisy con-
cert the patient enters a condition of ner vous exaltation. The crisis occurs,
rhe result of evident hypnotic suggestion. He rises and dances wildly in the
hut. The hubbub is redoubled. The y implor e the spirit to declare his
n ame. A name is shouted ... " (ibid., 443 ). This is very close to what
Zempleni describes as occu rring among the Wolof. In his account of a
long and theatrical ndbp ceremony he describes the musicians once more
beginning their drumming, the chants resuming, the bells ringing "louder
than ever." Fin ally, the sick man's trembling "g ives way to organized
movements." Thi s is the ritual ized crisis, followed by a long cry and then
th e declaration of the name "that the y were waiting for."
We have another example from T anzania. Am ong the Segeju (Gray
1969, 178 ) the sick man's public declaration of the name of the shetani
possessing him occurs at the climax of a seven-da y ceremony character-
ized by "noise, dancing, blood, and frenzy." This declaration takes place
on the sixth da y. On the seventh the spirit leaves the patient.
In the case of a quite different form of crisis, the wi ld possession crisis
mentioned earlier, what happens differs considerabl y since, in this case,
music is used not to provoke the crisis or fit bur, on the contrary, to make it
subside . At least, this is wha t Herskovits (19 38 2:186 - 87 ) found among
the Fon .
"On a given night the voduno informs each candidate that one of the vo -
dun of the pantheon will 'declare war' on him. " The cand idates are awak-
ened before dawn by a gun shot and loud cries. The y leap up "from their
sleeping m ats, run about wildly, throwing stones at anyone wi thin reach,
and beat ing those who approach them." A drummer w ill then succeed,
with some difficulties, to cal m them down, by beating out rhythms that
·'will bring them their gods.'' And it is at this point, after the name of the
divinity has been proclaimed, that the candidates, calmed by the drum, go
into trance. As among the Thonga and the Wolof, this episode is thus
linked to the declaration of the god's name. In the candomble, or at leas t
in the old candomble ritual described b y Cossard (1970, 204 ), which, as
\\·e ha ve seen, constitutes one aspect of wild possession-in this cult, events
unfold in \\·ays comparable to the example just described among the Fon .
The ere in the process of initiat ion were "rele ased into the countryside
where the y lived a solitary existence, hiding from the eyes of the lay people
and living on wild plants, or even on whatever they could steal from peo -
ple 's homes ." Summoned back afte r a period of several da ys, the ere
would return "in a state of extreme savagery, beating anyone they encoun -
tered on their path" (ibid. ). Possession of the ere by their respective orisha
was then induced by a specific drum rh ythm , the adarum (to which we
shall return ) thereby causing them to pass from a state of wild possession
to what we ma y call, antithetically, "cultivated" possession. Here aga in
Music and Possession 71
music clearly plays the role of calming the wild possession, not of provok -
ing it. "
ow let us examine possession when the people concerned are no longer
candidates or adepts in the course of initiation, but initiates. Here again
the relations between music and possession a ppe ar to be quite va riable.
Possession ceremonies a re organized in va ri ous ways depending on the rit-
ual to which they correspond: the coming out of ne w initiates, for exam-
ple, enthronement of a new dignitary, the annual feast of the divinity, sea-
sonal rituals, conjuration of calamities, therapeutic sessions, and so on;
the calendar and frequency of such ceremonies naturally vary a great deal
from cult to cult. In a general way, however, crises or fits and possession
trances do tend to occu r at certain spec ific points in the ritu al and not at
others. In the majority of cases the y occ ur while music is being played, and
are closely related to it. It is essential, all the same, to stress not only that
the rule is far from being absolute but also that the same music, or rather
the same type of music, does not have the same effect on all adepts, re -
gardless of circumstances. Thu s adepts in a state of impurity must never
under any circumstances be possessed. This means that the effect of music
is not as constrictive or as automatic as some people have sometimes
wished to believe. P. Simon and I. Simon-Barouh (1973, 42 n. ) provide a
well -substant ia ted example in their descr ipti on of the hail bong. But al -
though this restricti on on account of impurity, which concerns wome n
above all, seems to be unive rs al, there a re othe rs, on the contrary, that are
spec ific to certain cults .
Among the Yoruba the annual feasts for Shango, god of lightning, are
attended by a gre at number of adepts. All of these, and there may be sev-
eral dozen of them, are capable of being possessed by Sha n go. But in prac-
tice the god must never be embodied by more than one of them through-
out the entire feast, which lasts for seventeen days. 18 On the first day the
man chosen goes into his trance under the following circumstances: at a
given moment in the ceremony, soon after a kid has been sacrificed to the
god, the drums fall silent; an officiant then chants an invocation' 9 during
which his female counterpart holds out the an im al's freshly seve red head
toward those who a re present. Suddenly one of the adepts goes into a
rrance, dashes forward, and seizes the kid's head between his teeth, which
is a sign that the deity has accepted the sacrifice. From that point on the
man is possessed by Shango . He will remain in this state during the follow -
ing seventeen days, although he will alternate between the possession st ate
and the ere state described earlier. In the course of one season the feast
moves from village to village, and in each one a different adept will in turn
embod y Shango. The same invocation, chanted in the same context, thus
triggers the trance in this or th at person , leaving all 0thers unaffected. The
effectiveness of the music thus appears to be largely governed by co n ven-
 72 CHAPTER iWO
20
tion. And it should be remarked in passing that the convention is differ -
ent among the neighboring Fon and Gun. Among them, Khevioso, a
homologue of Shango, can be embodied by several dancers simultane -
ously. The same is true in the case of Sakpata , who admittedly has a great
many hypostases, wh ich is not the case w ith Shango.
As a general rule, it seems that whatever the cult concerned, the reac -
tion to music va ries according to the senio rit y and importance of the
adept. Wr itin g about the relations between trance and music, Cossard ob -
serves that "the more recent the initiation , and the stronge r the sensitiza-
tion to the rhythm, the greater the likelihood for the occurrence of trance"
( 1967, 17 5 ). In the ndbp, to which we will return late r, trance, under the
effect of mu sic, must normally result in the crisis and then the fall. Al-
though new adepts are "grounded by griots in a few seco nds" (Zempleni
1966, 416 ), the veterans, on the contrar y, do not succumb for quite a long
time, the actual period varying acco rdin g to the individual. Whe n, for
their part, women officiants ( those who o versee the cult ) go into trance,
the y are in theor y not supposed to go "as far as crisis and fall," unlike the
adepts alre ady mentioned. (ibid., 400,417 ). If it so happens th at one of
them does - and the honor in such a case goes to the first drum, who uses
ever y ounce of his talent to produce this end result - "the session then at-
tains its highest peak of intensity."
The efficac y of music thus va rie s according to the status of the adept,
who can be more or less vulnerable according to age, experience, an d po-
sition in the hierarch y of the cult. We shall return later to the role played
by the musician, as opposed to music itself, in inducing the trance state .
For the moment we can say that, regardless of the particular cult , the offi -
ciants, who by definition have had a long experience of possession, can
control their behav ior to a much grea ter extent th an new initiates. This
means that the y a re less slavis hl y controlled by the music, so that the y can
either resist it, as our ndbp example has just sho\,·n, or in fact do without
it altogether, as we learn from Michel Leiris 's descriptions (195 8, 61 - 71 )
i
of the possessions of Malkam Ayyahu, a pr actitioner of the zcir cult, who
could shO\.v off the possession state at the drop of a hat, as it were, outs ide
any rituall y defined occasion and also, needless to sa y, withou t any musi-
ca l stimulus. Generall y spe aking, moreo\·er , since the role of officiants is
to control the possession of others, it is natural enough they should first
know how to control their own .
These few examples will have made it clear enough, I trust, that the rel a-
tion s between music and possess ion va ry to a very great extent depending
on the particular cult, the stage reached by the adept, the state of that
adept at a particular moment, and the ritual in wh ich he o r she is taking
part. In certain cases, music tri gge rs the fit, in others, on the contrary, it
brings it to an end. Sometimes it seems to pla y a decisive role in inducing
trance; sometimes it seems to have none whatsoever.
Music and Possession 73

As for coming out of trance, as opposed to entering into it or being in it,

music can again play a role. In Bali, rhe trance of the little girl dance r s is
ended through specific songs (Belo 1960, 198 ). In Chad, among the Mun-
dang, in order to return the possessed persons to their normal state, the
musicians pla y a particular theme while the chief officiant taps them on
the back. 21 In the cand omble it is a rh ythmic theme pl ay ed by the drums
that tells the dancers embod ying the orishas to lea ve the room at the end of
the ceremon y (Cassard 197 0 , 11 7).

W H AT M SIC?

There are certa in ca ses , as w e hav e just seen, in which the triggering of cri -
sis or the entr y into trance can in no wa y be attr ibuted to the effects of mu -
sic, for the ver y good reason that at that p articular moment there is none.
This is exception al, howe ver. As a general rule , possession fir or trance is
accomp a nied b y music, and music is almost alway s reg arded as being
m o re o r less responsible for its onset. In w hat does this mu sic consist? A
swift glance at the evidence shows th at from the view point of its formal
char acteristic s it varie s considerabl y, if not totall y, from one countr y to
another. Ultimatel y, we hav e to accept that there are as man y different
kinds of possession music as there are different possession cults. Yet all
these different kinds of music a re regarded as being responsible for induc-
in g tr a nce . H ow can it be th at, despite their re specti ve differences, the y all
hav e the same effect? That is the question. The answer is that in order to
h av e the same effect - and a ver y specific effect, let us note - the y must
share , despite their differences, the same characteristics. We shall now at-
tempt to identif y these char acteristics by examining one by one the princi-
pal aspect s of these variou s kinds of music.

Vo cal o r lnstrument aP
Let us begin \\·ith the aco ustic ch a racteristic s o f tr ance mu sic. Of what is it
m ade up) ls pos sessio n musi c gener allv vocal o r instrument al , o r both at
t he sa m e tim e ) \"X
l hen it is instrum ent al, is o ne in strum ent mo re frequentl y
use d th an any other? In w hat cas e is a cert ain in strum ent used? Let me say
at o nce th at pos sessio n musi c is, as a gen eral rule, both voc al and instru-
ment al, but th at th e rel ati ve imp o rt a nce o f th e vo cal and instrumental ele-
ment s va rie s a great de al fro m o ne cu lt to anorher and tha t it also varies in
a ny given cult acco rdin g to how th e ritu al unfolds and ho w the tr ance de -
velop s . There are, ne verthele ss , sever al ex cepti ons , so me o f w hich I sh all
now cn e.
Th ere are ex amples of p os sessio n acco mp anied by singing alone. In
J ava, a possession cerem o ny centered around a doll , and whose principal
goal is to induce r ainfall , is ac comp anied by purel y vo cal music (Kartomi
74 CHAPTER TWO

19 73, 179-82 ). In Bali, the trances that Jane Belo labels as folk trances,
wh ich are due co possession by sanghyang spirits, not by gods, seem also _
to be acco mp an ied by purel y vocal music; at least Belo's descriptions
(196 0 , 201 - 25 ) do not mention anyth ing but songs." These tr ances are
ver y violent and usuall y end with a crisis and fall. We shall see in a later
chapter ,; that in Tuni sia, dhikr ch ant s' "- here purely a cappella - whose
primar y function is to accompany communion tr ance , ca n in a particular
context trigger possession tr ance, but in this case those who are affec ted
are not the ones who sing the chants but those (women, in the present
case ) who listen to them. Black Africa, coo, provides us with specific ex-
ampl es of purel y vocal possess ion music. I h av e already ment ioned one of
these a little earlier w ith reference co the orisha Shango. In the bori cult, at
the end of the no vice's purification ritual, the bori women who partic ipate
in th e ceremonies enter into trance when the y hear the two women offici -
ants singing the monos of their gods (Monfouga-Nicolas 1972, 109 ).
The se women can also become possessed under quite different circum-
stances, w hen the y hear these same monos sung by the ir handmaiden -
wives (their el-biya ), w ho, in truth, accomp any their singing wi th the rus-
tling of co w ries the y hold in thei r hands (ibid., 16 0 ff. ). In other words, all
these songs, contrar y to the usual practice in the bori, take pl ac e without
the presence of an instrumentalist. Among the Wolof, during the samp rit -
ual , which is a va riant of the ndop reduced to its "essent ial symbolic mo-
ment s" (Zempleni 1966, 426 ), the music consis ts solely of sung bak,
which are the monos of the gen ii ( rab ). As for the neoph yte 's possession, it
is reduced to a simple trance accompanied neither by d ancing , nor fit, nor
fall. The drummers, who pla y such a cruci al role in the ndop ceremonies,
do nor rake any part in the samp - a ritu al chat is "very strongl y associated
with the Moslem rab ." This absence of drums is explained in va r io us
ways, all connected with the fact that the samp appears to be "the princi-
pal means of effecting a compromise between Islam and the rab cult "
(ibi d., 428 ). This abse nc e must therefore be seen as connected w ith the
disapproval of mu sical instruments so ch arac teristic of Islam. Be char as it
ma y, it is nevertheless true th at w hether \\·e are dealing w ith the bori, the
rab or the orisha cults, such pu rely voca l mu sic mu st be seen as the excep-
tion in bl ack Afri ca, since the mu sic acco mp any ing pos session is generall y
pro vided , there as elsewhere, by a combination of instruments and voices.
Inversel y, we can also cite ra re examples of purel y instru mental posses-
sion mu sic. In Bali, in the barong, the music that accompa nie s the trance
during the final episode of the kriss dance is provided solely by gamelan in -
struments, and is not accompanied by any so ng s (Berthe 19 70, discogr a-
ph y ). In Brazil , in the candomble, there is a particular kind of drumming,
adarum, that ma y not ha ve any vocal accompaniment (Bastide 195 8, 20 ),
and wh ich is specificall y intended to provoke possession not in one par -
Music and Possession 75
ticular adept but in all the adepts at once. In ancient Greece, the dances of
the Corybantes were performed to the sou nd of purel y instrumental mu-
sic; this is at least what the writings of Plato and Aristotle indicate, as we
shall see later. " In so uthern Ital y, ta rantism would a t first sight seem to
constitute a read y-made example of purel y instrum ental p ossess ion music,
the tarantella being thi s cult' s central musical expression. This dance is
performed to music provided b y an instrumental.ensemble un accompan-
ied b y any so rt of sin ging. But if we look more closel y we notice that there
also exist "singin g" t a ra ntul as who require a voca l fo rm of "iatromusic"
(Carpitella 1966, 362 ), an d pizziche tarentate, 26 wh ich a re sung. In taran -
tism , song, compared to the tarentell a, h as little imp orta nce. It can, never -
thele ss, t a ke the most diverse forms, from stanzas ladden with erotic allu-
sion s to funer a l laments -the use of a particul a r type o f mu sic being
dict a ted b y the circumst a nces and being dependent upon the ritual and the
type of possession con cerned. It is in fact quite difficult , as mu st be cle a r by
now, to find p ossess io n ritu als in wh ich the music is always an d exclusi ve-
ly instrumental.
In the majorit y of cas es, as I ha\·e said , p ossessio n mu sic m a ke s use ,
either simult a neously or alte rn ately, of bo th song and ins truments. The
combinations of the two are quite va riable, howe ver, and it woul d seem
th at so ngs tend to be int errupted m o re frequentl y th an the instruments.
But what a re these instruments' Although the distinction m ay not al -
ways be valid, they ma y be assigned to one of two categories acco rding to
whether their function is melodic or rh ythmic, wi th those in the second
category most frequentl y serving to accompany th e first categor y of in-
strume nt s or the voice. Certain instruments, such as rattles o r be lls," h ave
a rh yt hmic function by their ver y nat ur e, as it we re, while others have a
naturall y melodic function , like the fiddle." Others, such as the drum,
can pe r form one or the other function depending on the \\·a y the y are
used . In reg ions where tonal languages a re spoken, the drum can be made
to "speak" in such a way as to perform rhe funcr;on of a melodic instru-
ment and acts as a substitute for so ng. Apart from this particula r case, it
may be pla yed in such an exp ressive manne r that the drumming can be-
come, as in the ndop, a melodic comb in at ion o f timbre, st re ss, and inten-
sity . In sho rt , if the drum is the instrument most often used fo r possession
m~s ic, this is because it can be melodic as well as rh yt hmic, a nd because,
\,·hen fulfilling the second of these functions, it can be incorporated into a
great man y different instrumental ensembles . 29
T his being so, if we look at the vari ous kinds of possession mu sic as a
w h ole, there does not see m to be one single type of melodic instrument
that noticeabl y pre va ils over the others . The violin, nowadays, is the pre-
ferred instrument in tarantism, but in earlier t im es its role could be filled
b y a gre at va riet y of instrum ents, amo ng them musettes, shawms, trum-
 76 CHAPTER TWO

pets, bombardons, and other wind instruments, which are sometimes re-
placed today by the accordion (Carpitella 1966, 361 ).
Piano has never been mentioned in this context, as far as I know, but it
has been in another, not so far away. In Paris, arou nd 1780, Mesmer, the
well -known inventor of magnetic trance (which h as thousands of adepts
nowadays in Latin America ) and also the sponsor of Mozart's Bastien and
Bastienn e, used to play "mo ving tunes" on this instrument to trigger the
trance of his patients when the y could not attain the "pe rfect crisis"
(Rausky 1977 , 104 ).
Leaving Europe now for Africa, we see that among the Songhay of Ni -
ger (Rouch 1960 ), the fiddle is the instrument used to play the spirits' mu-
sic . This is also true among the H ausa, though in certain cases the fiddle
ma y be replaced by the flute (Monfouga- icolas 1972, 105 ). Lee us also
observe only in passing - since the trance concerned is one of communion,
and not of possession-chat the flute plays an essential role in the orches-
tra used to accompany the Me vlevi whirling dervishes in their d anc ing and
trances. We should add, though, that there are many dervish sects and
chat the Mevle vi one is among the few chat uses che flute (nay )- an ob -
lique one, incidentall y- for the dhik r. In Iraq, co cite only one example,
che der vishes who practice "body beacing"-in ocher words corporeal mu-
tilation-during the dhikr do so to the sound of an orches tra made up
solel y of drums, with no accompanying melodic instruments (Hassan
197 5, 223-2 6 ). Bue in the d omain with which we are dealing here, trul y
ever y rule is immediatel y contradicted by a counterexample.
In ancient Greece the double clarinet (aulos ) was the chosen instrument
for the Dion ysiac rituals during which the maenads and Corybantes sur -
rendered themsel ves to the manifestation of their enthusiasm or mania. 30
As soon as the y heard it , th ey fell "into the trance of divine possession"
(Jeanm aire 1951, 175 ). But although there are conclusi ve rexes (notably
Plato 's Symp osium ) chat connect the triggering of tr ance with the sound of
I the aulos, and although both the maenads and sat yrs are often repre sented
\:
' '
i.:
as blo w ing into twin clarinet-like instrument s, the Pan flute (syrin x ) is also
i'
the instrument o f the sileni (Sympos i um 21 5 b ) and thus of the musicians
of Dion ysiac "enthusiasm. " The maenad s are sometimes depic ted as strik -
ing tambourine s in order to induce other entr anced maenads to dance.
One vase shows Dionysus himself pla ying the seven-stringed lyre 31 in a
scene of bacchic dance (Emmanuel 1914, 397 ). Another vase shows a sa -
tyr also pl ay ing a lyre , but a five-stringed one this time (ibid. , 395,396 ),
while a d ancing maenad executes "swa ying mo vements of the torso, arms
out stretched, " in a pose that Jean Rouch 's films on possession among the
Songha y have made familiar to us. True, it is clear from the rex es chat the
aulos was the most typical instrument of corybantism and maenadism,
but it is also true, as we have seen, chat the Greeks assoc iated ocher instru-
Music and Possession 77
ments w ith trance, up to and in cluding the lyre itself, whic h since it was
Apollo's instrument, one might have expected to be excluded from Dion y-
siac rites, and w hich, in Crete, was associated w ith the aulos in the ecstatic
dances of the kouroi (Jea nm aire 1939 , 432 ).
In Tibet , just as the alre ady entranced di viner ' 2 is about to be totally
possessed, "two pr iests approach the o racle and tak ing up po sitions at his
left and right sides, eac h of them blows a thi ghbone trumpet into h is ears"
(Nebsky -Wojkowitz 1956, 430 ). W hen he he ars this sound "the intensit y
of hi s convulsions incre ases." Among the Sum eri ans it was the harp that
was "used during the oracles of the high priest " (Farmer 1957, 235 ).
Among the Hebre ws (and we will return to them later ), st ringed instru-
ments , the harp and the lyre, we re asso ciated w ith prophetic tr ances. But
in certain cases the stringed instruments we re complemented by clarinets
and drums. 33 In Ir aq (H assa n 197 5, 181 ), the lyre is the typ ical in strument
of possession cults practiced by the blacks originating from Afr ica. In
Chad, am o ng the Kotoko (Br andil y 1967, 53 ff. ), the characteristic instru-
ment of p ossess ion rites is the harp, accompanied by t he wate r drum;
among the Gula Iro it is again th e harp, but even mo re pr evale nt is a series
of spherical wh istle s (Pairault 1966, 25 9 and fig. 179 ); among the Mun-
dang (Adler and Zempl en i 1972 ) it is the go urd trump et , and among the
Mussey (Carine, film og r aph y) it is the gourd trumpet w ith vibrator. In the
Ce ntr al African Republic, among the Ngbaka -Mandja (Arom a nd T aur -
elle 1968 ), it is the xylophone. In Zimbabwe, amo ng the Shon a, it is the
mbira, a set of plucked lamellae (McEwen 19 70 and Tr acey 1970, 37-
38 ). In Madagascar, among the Vezo (Koec hl in 197 5 ) it is the acco rdion ,
wh ich has rep laced th e zith er for thi s purpose . In Bali, unlike the "fo lk "
possessions menti oned a sho rt w hile ago, wh ich are acco mp anied by purel y
vocal music, possessions attri buted to the gods a re accompan ied by gam -
elan music (or rather, to use the Balinese wo rd , by "gongs" ), some time s in
asso ciation wit h song and so m etimes no t. In Laos (J . Brunet, discogra -
phy ), mouth organs accompan ied by small cymba ls are used "to m ake the
shamans go into trance. " In Vietnam, possession songs ate normally ac -
compan ied b y a moon -shaped lute and a little drum. 3 '
Wha t can we deduce from this ser ies of examples - whi ch could easily
be extende d although multiplying them would only serve to increas e the
variety of cases - excep t for the fac t tha t no constant s emerge from it, or ,
more correctly, t hat rhe one thing th at does emerge is the lack of con -
sta nts? Th e conclusion , in co nsequ ence, is that among all the instruments
used fo r possession music not o ne stands o ur as the rule. Practicall y spe ak -
ing, they can all do the job; at least, th is is what the facts indicate. One
may wonder if so m e particular instrument, by virtue of it s specific
sound - vio lent , strident , piercing , scre echin g, en veloping, haunting, per -
cussive, who knows? - might perh aps have turned our to be more capable
78 CHAPTER TWO

than any other of contributing physically, in one way or another, to the

preparation for and the triggering of trance. Clearly this is not the case at
all. If there were such an instrument, then one would come across it, if not
invariably, at least more frequentl y than any other. The facts indicate the
contrary. One may also wonder that, despite their great variety, the musi-
cal instruments used for possess ion do not manifest some commonly
shared acoustic property that m ight perhaps explain why they all act upo n
the listener in the same way. But no family of instruments is missing from
our list above; the hypothesis is thus untenable . Drums of all sorts - and
heaven only knows that their variet y of forms, dimension, materials, and
sonorities could hardly be greater-bells, rattles, harps, lyres, fiddles, xy -
lophones, gongs, crumpets, clarinets, mouth organs, whistles ... Every
o ne of these instruments, wh ich somewhere or other is closely associated
with tr ance, delivers acoustic signals too radically different from one an -
other to be able to produce, at the level of audirory physiology, the same
effects on the listener. So that if one of then could by chance produce such
effects - that is, were capable of triggering trance by virtue of its sonority
alone - anothe r would certainly lack this ability. Although all these instru -
ments are associated with trance, and more particula'rl y w ith entry inro
trance, this association is not related t0 the ph ysical impact created by the
sounds they produce . At the beginning of this chapter we asked ourselves
wha t particular characteristics in struments used for possession could
share? We now see they have none .

Rhythmic, Dynamic, Melodic?

Music, we have seen, can be vocal or instrument al, but whether it is one
or the oche r still does not explain the sort of action it might have on
trance . Its effects cannot be attributed to some special virtu e of the vo ice ,
since in trance sess ions there might quite often be no singing at all, nor can
it be attributed tO any sui generis power of a particular instrument sinc e,
in fact, all instruments can be used, contrary t0 the commonly accepted
idea th at the drum is the major cause of the great hurly -burly of posses-
sio n. We must now investigate if the effects of music could be due not tO its
aco ustic characteristics, but to the manner in which they are shaped; by
this I mean by the rh ythmic, dynamic, or melodic - o r more correctly
modal - features of the music.
Are there such rh ythms that in and of themselves, because of their in-
trinsic musical structure , are more suitable th an ochers for inducing
trance? Concretely (and presenting the facts in a r.acher simplistic way ),
would 4 / 4 time be more likely to produce such an effect than 3 / 4 time? A
syncopated rh ythm more than one that is not? A staccaro more than a le-
garo rh ythm? A rh ythm with strong stresses more than one almost w ith -
out stress) A monoronous rh ythm more than a continually cha nging one?
Music and Possession 79
Carpitella (1966, 361 ), who considers "rhythm to be the determini n g
facto r in tarantism's iatromusic," asks if the original rh yth m of the tara n-
tell a was in '"pe rfect ' or 'imperfect ' time." Danielou, for hi s part (1975,
14 ), asserts th at "t he processes and the characteristics of these [ecstatic]
dances a re the sa me in all regions of the globe , in Indi a, in the Middle
East, in Africa," and th at "the rh ythm s used are always of 5, 7 or 11
beats." Such an assertion can not be taken seriously: no study h as yet been
undertaken th at would allow us to sta te th is uni versal a rule, an d a great
man y obse rved facts contradict it. T he question posed with respect to the
tarantella is, however, legitimate, and we are justified in asking the sa me
sort of questions about an y music used to accompa n y trance dances.
As far as I know, no work, either monographic or comparati ve , ha s
been publi shed on this subject. H owev er , it seems va lid to ask, for exam-
ple, what exactly were those rh ythm s am ong the Greeks "th at h ad the
property of provoking trances, by possession of the Cor ybame s and also,
no doubt, of the Sileni-Sat yrs" (Jeanm aire 1951, 49 7 ). Plato himself
scarcel y had an y opinion on this point. In Th e Rep ublic (400 a- b ), wh en
discussing what I shall label for simplic ity's sake th e "ethos of rh yth m s," he
adm its he does not know either "which rh ythm represent s such and such a
character," or "which rh ythms are suitable (fo r producing ) m ania, " in oth-
er words tr ance . We know from other source s that the dithyramb was
originall y closel y associated with Dionysiac rituals. And as a poetic genre,
we are told th at it was essentiall y iambic, so its rh ythm must ha ve consist -
ed of a foo t made up of on e short syllable followed by a long one. In a text
att ributed t o Plutarch and presented by Jeanmaire ( 19 51, 289 ) as "a devel-
opment that is basicall y no more than a paraphrase o f the Pia conic theor y
concerning the divine origin and char acter of m ania," we read, conce rnin g
the relations of possession to rh ythm and melody, that "in the case of Bac-
chic and Cor ybantic leaps, it suffices tO change the rh ythm by aba ndoning
the tro chee , and the air by abandoning the Phrygian mode , for them to be-
come less vio lent and cease " (Amatorius 16. 3 1). He re, then, the trochee
(a long syllable followed by a short ), and not the iamb , is associate d with
possession dance. El sew here , apropos the wo rd iamb os, Jeanmaire
(1951, 234 ) wonders whe ther it refers "in one case to a dance step in duple
time and in the other to a rh ythm that ma y be in triple tim e. " Unce rt aint y
on this point is evidentl y considerable . What we do know, on the othe r
hand (ibid., 242 ), is that "the celebration of the dith yramb manifested it-
self by nois y exclamations, outcries under scored by music in wh ich the
flute [read: aulos ], which incites mani a, pla yed a role that has always re-
mained preponderant, but w hi ch neve rth eless included other instruments,
small drums, kettle drums , castanets. The rh ythm must have been d omi-
nated b y rep eated ritual acclamations." Int eresting though it is, this text
does not tell us much about exactly what this rh ythm was (a po int to
which we shall return late r, but from a different angle ).
80 CHAPTER TWO

Just as the lyre, Apollo 's instrument , is sometimes assoc iated with
Greek Dionysiac cults, so in Eg ypt the paean , an essentiall y Apollonian
literary form as well, seems to have been incorporated, with its own spe -
cific merer no doubt, into the Greek Dion ysiac tradition. 31 Such is the in -
formation we h ave concerning the rh ythm acco mp anying tr ance and to
wh ich Dion ysiac dances we re performed. As we can see, it is r ather scant,
but there is enough to let us suspect that no one rh ythm was especiall y
assoc iated among the Greeks with po ssession d ance, or, to say the same
thing differentl y, that there m ust have been several rh ythms, and that none
was specific to tr ance. As a m atter of fact, thi s is the same conclus ion we
shall re ach again later on, albeit by way of d ance , in the chapter devoted
to possession music among the Greeks.
From an cient Greece let us no w p ass on to modern -da y Brazil and t he
mu sic in candomble, about w hich we do possess a fair amount of informa -
tion , owing in particular to the work of Gisele Cassard. T he repe rt oire of
twe nt y-five rh ythmic themes published in her study (1967 , 180 - 207 ) of
candomble music contains, with only one exception, duple meter
rh ythms, in m any cas es empl oy ing triple subdivisions. As for the tempo
indic ate d, the y show that rhe smallest pulse unit va ries according to the
theme from MM 288 to MM 576 . 36 Thus we are faced with a group of
rhythmic theme s that are extremely uniform as reg ard s meter , but ex -
tremel y va ried as reg ard s temp o . Apart from one o r two , including the
one fo r the entry of the orlsha , the se theme are all intended to trigger
tr ances . Each of them is specific to a particul a r orlsha, w ith the exception
of the ada rum rh yt hm ,3' whi ch we al ready mentioned, an d is valid for all
the orlsha, wi th ou t any di stinctio ns. "Thi s sw ift and insi stent rh ythm is a
verit able imper ative," says Cossa rd ( 1967, 176 ). "Ra re are the initiate s
,, ·ho can re sist it, even chose among the most senio r." On the da y of the
center's ann u al closure, it is chis them e char the saint's father orders to be
pla yed "so that all the initiates present ,,·ill go into trance " (1970, 145 ). It
is also played during any oth er cerem ony when the orlsha "dela y from in-
co rp orating them selves" (ibid . , 108 ).
In fact, the adarum does not have a pa rti cularly quick temp o, since Cas-
sa rd records it as being MM 69 to th e quarter note . It contains no note
sho rt er chan a sixteenth note and frequently makes use of both eigh th and
qua rter notes. Its rh ythm is nevertheless very particular. The first three
bars co nt a in accented and unaccented eighth notes, an eighth note against
the beat , and a trip let of eighth and sixteent h notes; the next three bars are
made up , on the contrary, of a regul ar sequence of quarter notes. Rh yth -
mically, then, thi s theme has an extremely uneven profile. As for its "insi s-
tent" character, it evident ly ste ms to a large extent from the four acc ented
qu avers wit h wh ich it begins. ls the efficac y of the ad arum due to rh ythmic
factors comparable in so me ,,·ay to th ose Metraux (1958, 161, 169 ) de-
Music and Possession 81
scribes in Haiti when he wri tes about the drummer's "feims," defined as
"breaks" against the beat ("brisures et cassees" ) that "interrupt the flow of
the dance and produce a state of paroxysm propitious for loa crises?" Does
this device belong to a musical system specific to tr ance? Are not the same
devices equally used in music having no connection to possession? An in-
depth study would be needed co ascertain it.
However, let us note that evidence of the same kind has been collected
by Alain Danielou (1967, 92 ) in Asia, "in the kfrtana -s, wh ich in Bengal
are the mystic chants danced to in groups, and in the prophetic dances of
primitive peoples in Southern India." Here is what he writes: "The dancers
are first of all drawn into an easy rhythm wi th w hich they identif y com-
pletel y and thus sink into a sort of hypnotic half -sleep. The musicians then
create a shock by means of several violent drum strokes and embark on a
new, much more complex rhythm. After hesitating an inst ant, the dancers
are taken over by this new rhythm ·without even consciously willing it. In
some of them, this provokes a trance state and a complete loss of self con -
trol, as though the rhythm were a kind of spirit that h ad possessed them.
Thi s trance state is characterized by insensitivity to pain, complete loss of
modesty, and visionary perceptions."
Let us also remember, even though it lies somewhat outside our domain
since it involves shamanism and not possession, that according to F. Den s-
more ( 1948, 3 6- 3 7 ) the healing songs of the orrh American Indi ans she
studied are characterized by changes of stress and rh ythmic al irregulari-
ties th at distinguish them quire clearly from the tribes' other songs .
Although there are not enough data at our disposal to enable us to state
that such rh ythmic breaks const itut e one of the universal features of po s-
session music, it is nevertheless true that the y recur very frequently, and
this is something we need to remember. On the contrary, it does seem that
we can consider another rhythmic feature, the acceleration of tempo, co
be universally used as a means of triggering trance. Discussing the dirh y-
ramb as a ritual, which is co say as a "chorus arr a:1ged in a circle that sings
and dances," Jeanmaire (1951, 240 ) write s: "Th e mo vement was imparted
only gradually to rhe cyclical chorus, which for a w hile at leas t gave the
impression of crawling on the ground like a hea vy cable , the goal being a
progressi ve approach co the ecstatic state, which is achieved precisely by
the monoton y of the acclamations, repeated at first co a deliberately slow
rh ythm , and by th e progressive accele ration of their cadence . " 38
Writing about rhe candomble, more specifically in reference co the toque
adarum and after having stressed that "it is not accompanied by chants,
since this time not just one divinity but all of them at once are summoned,"
Roger Bastide ( 19 5 8, 20 ) notes th at "the ever more rapid rhythm, 39 the
ever more imploring tone finall y op en s up their muscles, their viscera,
their heads, to the penetration of the god they have awaited so long." Simi-
Music and Possession 83
tensified by a "very pronounced sforzando when the chorus took up the
dominant phrase" (ibid . , 442 ).
In the candomble, "when possession is long in coming," Roger Bastide
wri res (19 5 8, 18 ), "the priest or the priestesses ring the agogo 41 close to
the girls' ears as they dance, and then it is by no means rare for the divinity,
spurred on by this deafening noise, this barbaric sound, to make up hi s
mind at last to mount his horse." Here the crescendo is achieved not only
by louder playing but also by playing closer to the one for whom the music
is destined. The same is true, as we ha ve just seen, w ith the Thonga and
their rattles. In Tibet, as we saw, horns are blown into the ears of the ora-
cle on the verge of possession . In Bali, Jane Belo ( 1960, 50 ) describes how
the singers leave their places and mo ve to a spot just behind the medium
who is about to be possessed, so that they can sing very close to him and
thus make him go into a trance. These facts are nor w ithout importance,
and we shall retu rn to them later.
In Madagascar, among the Vezo, B. Koechlin (see discography ) notes
that when the god rakes a long rime ro come down, rhe tempo increases,
while at the same time "interjections ring out : 'yo ur hands!,' 'louder,' 'clap
all of yo u"'; and among these exclamations , he selects "it's getting hot
now, sister!" Zempleni, for his pan, has demonstrated rhe great impor -
tance of the "warming-up" process in ndop ceremonies (1966 , 395,399 ),
which follow a carefully staged dramatic scenario. All the participants -
officiants, musicians, and spectators - contribute in varyi ng degrees to the
general warm-up. Although the chanting plays an important part in this
process, the decisive role seems to be pla yed by the drumme rs, the
"griors" - who themsel ves are well awa re of this fact: "We summon an d
excite the rab with our drums," they say (ibid., 395 ). The lead drummer
seems to be the "master in these rh ythmic dialogues between the public
dancing and the crisis." With the first "crisis labor,'' the bear of rhe drums
speeds up, then a little later, as the y announce and at the same time cause
the possessee's collapse, "the movements and the drum become very fast ."
The possession seance for rhe "naming of the rab" (th at is, for identifying
the sp irit possessing the patient ) as descr ibed by Zempleni (ib id., 35 6- 60,
369 - 70 ) becomes one immense and increasingly dramatic crescendo, in
which the drama (the fit and final collapse ) is carried to its climax by a
constant and progressive ampl ificati on of the sound. His accou nt is lit-
tered with observations relating to music. After a certain number of rit -
uals have unfolded, first withou t any musical accompaniment an d then to
the sound of a rattle (a small horn filled with cowrie shells ), an officiant
begins to ring "a little bell close to the patient's ears ." A short while later,
with the bell still tinkling, "the drums thunder our . Th e atmos phere in the
roo m changes totally." The ritual continues as the officiant "shakes his lit-
tle bell tirelessly in the patient's ea rs. The din grows louder still." The
84 CHAPTER TWO

treatment of the sick person takes eve r m ore violent forms, and "the dry
clicking of gourd rattle s is no w adde d to the al m ost intolerable noise of the
drums and the ins istent tinkling of bells." The sick wo man howls. She is ·
abou t to sp eak the n ame of th e rab. Silence then dr amatizes the situation:
"t he drums abruptly fall silent." If no answe r comes , "the d rums , bells,
and chanting resume." The acceleration of che tempo is de scribed else-
w here (p. 416 ) in these terms: "The drum beats and th e gestures then be-
come so rapid chat the ear percei ves a gualicaci ve cha nge in the music (even
though the rh yt hmic pattern stays the sa me)." The collaps e, "the termin al
po int of the cr isis," is "marked by a re sou nding blow on the first drum. "
Th e ndop, as described by Z empleni, is prob ab ly the m ost spectacular ex-
ample of incr easi ng dr amatic tension by an acceleration of mo mentum
and an intensification of sound. Bue degree of intensit y asi de, other exam -
ples of accelerando and crescendo abound, as we ha ve seen , and the list
could be extended wi th out difficulty - so much so that these two features
ma y w ell be regarded as constituting veritable uni versals of po ssessio n
music. Yee he re aga in several reserva ti ons must be m ade.
In rhe first place, it would seem char thi s dr am ati zation of the music is to
be found solel y in the possession cu lts in w hich trance is accompanied by
crisis, whe ther the latter pre cedes the tran ce (as w ith the Shan go ceremony
cited earlie r) or, on the contrary, forms its climax (as in the case of the
ndop ). W hen the tr anc e is totall y non convul sive in character, and w hen
the re is no fit, then the musi c m aint ains the sa me inten sity and tempo. At
leas t thi s is the case among the Mundang, as the films and rec ordings of
Zemple ni , Pineau, an d Adler demonstrate . ' 2 In this case, the dances per-
fo rm ed by the poss essees are always char acterized by a gre at gentleness
(even t he p an ther dance, so very different from w hat one obse rves during
the ceremonies for Age, god o f the hunt, in Abomey ). The music used to
acco mpan y chem - double- sk inned drum , gourd horn alt ern ately blow n
and sung int o, and rattle-is likewise devoid of all violence. It flows gend y,
always yielding the same deli very and it renews itself in a m anner chat
owes nothin g to changes of tempo or inten sity, bu t results from expressive
modulat io ns always occur rin g on the same registers . The charac ter of the
music, the character of the dances , and the charac ter of the trance thus ap-
pear to be completely interdependent . And nothing, indeed, could be
more natu ral.
Although the violence of music and of trance m ay well be closel y allied,
chis does not change the fact that an outburst of music may perfectl y well
not lead to a fit, or, in versely, that a fit can occ ur without mu sical vio -
lence. A.lchough the rel ation between musical violence and the onset of a
crisis seems to be extremel y close in the ndop, in other cults it appea rs to
be mu ch less so . True , in H aitian voo d oo, as in the candomble , to cite
o nl y two examples, trances most often occur at the moment whe n emo-
Music and Posse ssion 85
tions are at their str ongest and the music is at its most dramatic, but it is
not exceptional for things to happen quire differentl y. Michel Leiris told
me recentl y he had noticed this several times, as did Alfred Merraux, dur-
ing rh e voodoo ceremon ies th ey attended together. Ir is by no means rare,
they both obser ved, for trances to occur precisely at a moment when the
music slackens. 1o one would think of relating such slacken ing to the en-
tr y into trance. Yet the oppos ite seems quite normal, and if tr ance occurs
at the moment w hen music re aches its climax, no one fails to connect the
two. This is what has usuall y happened, and it explains why music has so
often been perceived, quite improperl y, as the very springboard for ent ry
into trance . Michel Leiris, for example, remember s a vo odoo ceremon y in
Haiti toward the end of w hich, at a moment when the music had ceased
completely, a lemo n ade seller, who was of course at a slight distance from
the d ancing area and thu s from the musicians, was suddenly seized by a
vio lent crisis and began rolling frenziedly in the dust and howling . And
Pierre Verger, too, has told me ho w often it happens in the candomb le that
whe re as the drummers ha ve unsuccessfull y attempted to send someone
into trance by unleashing the full force of their instruments , the trance is
brought on by the mere tinkling of a little bell gentl y shake n close to the
subject's ear!
In fact, we see that in the dom ain of the relati ons bet we en music and
trance , the connection bet w een causes and effects offers a ver y wide mar -
gin of freedom. To illustr ate this p oint, I will cite as examples two ceremo -
nies for Ogun, god of iron, that I attended wit h Pierre Verger in the Yoru-
ba region of D ahome y. In both cases the ceremon y was held on a weekly
basis in honor o f thi s orisha, and during it the officiants we nt into tr ance
alm ost as a ru le . The mu sic and drumming were quite spectacular (Verger
19 57, 150- 70 ). And as the recording s make pl ain: 3 rhe music attains the
sa me paroxys m al climax o n both oc cas ions. Yet on the first occasion the
expec ted tr ance s en su ed and o n the second the y did nor. There must un-
do ubt ed ly have been a re aso n for thi s; bur rhe fact rem ains that the sa me
music did n or produce the same effects o n bo th occas ions . I can cite yet an-
ot her exa mple of entr y into tr ance, which I w itnessed personall y, ag ain in
Dahomey, that was not linked to any kind of paroxysm . In this case (dur-
ing a ceremon y for Dan , the snake vodun ) it occurred w hile the offic iant
was chanting rather quietl y and shaking a rattle smoothl y and evenly. The
woma n, who was quire close to the offic iant, it is true , went into a tr ance
w ithin a m atter of minutes. As for East Afric a , K. P. Wachsmann (1957,
8 ) underscores the m any different aspects posses sion mu sic can assume
there. Although certain possessions are triggered by the sound of an ex-
tremel y nois y gourd rattle , others, on the contr ary, are induced by the
sound of ver y gentle chanting accomp anied by an almost in audible zither.
Then there is the extreme case of the "deaf " tarantul as (Carpitella 1966,
'I
1 86 CHAPTER TWO

362 ) who inspire in their devotees a beha vio r consist ing in "a nonchalant
dance, as though the tarantulee were deaf co the music ."
Investig ation of a number of different possession cults therefore show~
that, although the dramatiz ation of the music by accelerando and crescen -
do often plays a role of primary importance in trigger ing trance and/ or
fit, this rule is nevertheless far from absolute. Even in cults that make sys-
tematic use of these techniques, trance and fit can occur in the absence of
any musical paroxysm. Frequent though its use may be, this dramatiza-
tion cannot therefore be seen, contrar y ro what people often tend co think,
as the quasi-inevitable means of inducing possession. Moreover, such
techniques are in no way spe cific co possession. As far as I know, Ra vel's
Bolero, which is one enormous crescendo , does not habitually induce pos-
session in our concert halls. In Africa, an outburst of music, p articul arly
drum music, can reach the intensity described for the ndbp in the course of
feasts or ceremonies that have nothing ro do w ith possession.
Yes, one may say, but what about the ndbp, all the same? Is not the vio-
lence of the music in this case, if we are co believe Zempleni, the sine qua
non of the fit and fall? Let us leave aside the fact, significant th ough it may
be, that agitation and collapse also form pare of "nonrirualized" crisis
(Zempleni 1966, 314 ) which can occur, as we saw earlier, outside any rit-
ual framework and therefore in the coral abse nce of music. In "ritualized
crisis," in other words during a ndop ceremony, it is the crisis and fall, that
is ro sa y, the frenzied and convulsive dance and the fall rhar follows it, that
seem ro be directly associated w ith the explosions of the drummers, and
nor the entry inco tr ance . Moreover, the crisis and the fall are the norm al
(or even required ) end products of trance only for the neophytes. Experi-
enced possessees go into trance and dance, bur do nor go as far as a crisis
or, needless co say, a fall. Bur in what does this crisis or, co use the aut ho r's
own terms, this "labor of crisis" consist? Zempleni describes it as a "moco r
sequence" containing two consecutive phases, during which the possessee,
!, who is now "at the drummers' mercy," executes an increasing ly viole nt
x
I and convulsive dance. Th e second phase consists in "a rapid extension and
,,
il
bending of the arms, in an oblique line ro the axis of the body, and a vio -
lent swinging of the head up and down." The drummers accompany these
mo vements , "causing them co become still more rapid and violent all the
while ." Then comes a phase of disordered movements, contortions, rau-
cous cries, and then the fall. "The physiological cause of the fall," Zem-
pleni writes (ibid . , 417 ), seems co be aucosrimulation of the inner ear:
owing co the incre ase in volume and pace of the music, the possessee is lee
ro- or , more exactl y, dri ves himself coward - muscular exhaustion anc
extreme spatial disorientation. Thi s seems co correlate with the remarki
of Dr. H . Aubin quoted by Leiri s (1958, 18- 19 ) on the "frenz ied mo ve-
ments of the head ... subjected co repeated shaking in every direction,'
Music and Possession 87
on "t he violent excitation of the laby rinth" engendering a "special state of
drunkeness," and on the shaking "to which the thoraco-abdomino -pelvic
viscera in vol ved in some of the previously mentioned m ove ments" are
sub jected "to a lesser degree." It is true th at this frenz y is governed by and
reli es upon the drumming (even though for their p art the drummers, as we
sha ll see, p ay the closest attention to the beh av ior of the possessee ), but
the po ssess ee's loss of b alance and fall are due to the frenetic mo vements
of his dance and his own frenzied agitation in gene ra l, and not to the
music. Or, rather , the music is onl y indirectl y re spo nsib le fo r them; it
spurs his da ncin g on, bur it is not a source of vertigo or au ditor y diz ziness ,
as it were.
Wi thout ever explicitl y formulating t he h ypothesis , nu m ero us au th o rs ,
in rem arks scattered throughout their wo rk, more or less insid iou sly lead
the re ad er to belie ve that the violence of the so un d invo lved - particularly
that of drums, rattles, and bells - is capa ble of producing a phe nomenon
of thi s nature. "A din growing ever loude r ," "al most intolerable no ise of
the drums" (Zempleni ), "fr ightful hubbub" (Ju nod 1913, 441 ), "deafen-
ing noise" (Bastide ). We have al re ady encounte red "voo d oo ish drums of
the rad a ritual, unbelievably percu sive. . . . Orche str al convulsions.
Thunder " (Leiris 1955 , 18). One could add o ther quotations that hi gh -
light th e unbearable, int ole rable int ensity of sound ro wh ich the indi vidual
who is either posse ssed or on the way to becom ing so is subjecte d . Going
from these obse rva ti ons to thinking, or causing others to chink - even if
involu nt aril y- that an y convulsions, cr isis , o r collapse are in fact signs
chat th e intensity of sou nd can no longer be tolerated, involves o nly one
step . Thi s step, moreover, is all th e more easy to take because it is so me -
times taken (and for good reaso ns , at the level of 1ppearances at least ).
Describing the first possession crisis of che sai nt 's daughte r-to-be, an event
usually tak ing place during a feast, "when t he orisha are embodied and the
drums play the rhythmic compositions int ended to summon them ," Gi sele
Cassard (1970, 158 - 59 ) w rite s: "The perso n seems at first una ble to go
on tolerating the sounds she is hearing. She puts her hands ove r her ea rs as
though to protect herself from them, she sways, she loses her bal-
ance ... ." And this person, it should be noted, is onl y a mere onl ooke r ;
she is not part of the group of "e mb od ied orisha"; she is not dancing. So in
this case the gidd ine ss and the colla p se are to be ascribed di rectl y to the
mus ic, which "she seems unable to go on tolera ting. " Thu s we are faced
here wirh a very different process from the one obser ved in the ndbp.
But colla p se following a convulsive possession crisis can also be found
w ith out acco mp any ing mu sic . In Tibet , the medium goes into trance, as
we have seen, as a re sult of "special pr aye rs" chanted "in a peculia r, quick
rhythm" ( ebesky -Wojkow itz 1956, 547 ), accomp anied , depend ing on
the case, by drums , bells, cymbals, and sometimes thi ghbo ne trumpet s.
88 CHAPTER TWO

He is seated (ibid., 433-34 ), but as he enters his trance he begins "to swing
the upper part of his bod y rhythmically with a rotating mo vement from
left to right ... rolling at times limpl y as if it we re only loosel y attached
to the trunk." After having delivered the oracles demanded of him -
and after the music has stopped, of course - the medium is once more
seized by con vulsions, again rotates his trunk as at the beginning of his
trance, and suddenl y falls backward into the arms of his helpers, stationed
behind him to break his fall. The text does not specify it , but one assumes
th at the rotation of the torso is also accompanied on this seco nd occasion
by the headrolling mentioned earlier. So here we find the same bodil y
mo vements as those associated with the final collapse in the ndbp. The
ph ysical context and the relations to the music, howe ver, are very differ-
ent. There is no deafening music or , in fact, as in the second epi sode, no
music at all. There is no dancing properl y speaking, or any frenzied ag ita-
tion, except for a very brief period at the beginning of the trance . But the
medium does displa y an extreme state of inner concentration, achieved
through meditation that sometimes lasts for several days (ibid., 440 ) and
through a w ho lly corporeal technique that is not described but th at in-
cludes the swaying and rotation of torso and head alread y mentioned,
bre ath control, and an exercise that prepares the medium to bear the
weight of an ex tremel y heavy met al helmet, wh ich is placed on his he ad
as soon as the first signs of trance man ifest themsel ves and which he
wears throughout the seance. At cert ain moments the trance nevertheless
evidences very marked convuls ive aspects. Like the priests of Ogun in
Dahome y," the seated oracle suddenly leaps "several feet" into the air at
the beginning of his tr ance (ibid ., 434 ). Moreo ver, as in Bali and
Madagascar , his entry into tr ance is encour aged by the burning of a great
deal of incense . What role does this incen se play exactl y? Is it simpl y part
of wha t one might call an overall program of overstimulation? Or does it
in fact act as a true drug? The authentic oracles all abstain from an y stimu-
lant (ibid. , 440 ) as a means of entering the trance state . But there are "im-
postors " who make use of a mixture of Ind ian hemp and Guinea pepper
that "is supposed to m ake one feel the blood running hot through the
veins and become rather excitable by the music made at the beginning of
the ceremony" (ibid .). Clearly an intens ification of musical perception is
consciously sought here , wh ich me ans th at it must pla y some role in the
system. It is wo rth noting th at in Niger too, in the bori, during the cere -
mon y for the "raising of the dark gods," the neoph yte is like wise sub-
jected to fumig ation , and J. Monfouga - icolas tells us (19 72, 116 ) that
"just like the music ... the smoke from certain plants is a summons to
the gods, a sign the y recogni ze ju st as well as their motto ." In Niger , ho w -
ever, the plant s utilized produce an "acrid and evil-smelling" smoke that
ch ok es the initiate. In other w ords the stimulat ion sought is not at all the
Music and Possession 89
same as that in Tibet, but its physical effect is accompanied by a symbol ic
action that figures in a system of signs which here again is related to
music.
Let us return tO the final collapse of the Tibetan oracle and t0 the col-
lapse of the possessee in ndbp . The y each occur in contexts as different
from one another as could be imagined . In ndop, the collapse is very close-
ly connected tO the frenzy of the dance and the music; in the case of the
oracle, it is not. In the first case it is taken to represent "the lethal outcome
of the encounter with the double," writes Zempleni (1966, 419 ), who
then adds: "in order to let oneself be invested by, then fused with, one's
double [the rab], one must erase the foundations of one's own identit y,
one must die." In the second case, on the contrary, the collapse represents
"t he moment in which the dharmapala [deit y] is supposed to have left the
bod y" (Nebesky-Wojkowitz 1956, 429 ). However, it is true that in both
cases the collapse occurs after a period of intense emotion, extreme ten -
sion, an d gre at convulsi ve muscular exertions. It is probable, therefore,
that in both cases the fall is simply, in part at least, the effect of a brusque
decompensation. Howe ver, on the level both of ph ysical technique and
symb ol ic representation, the conditioning involved is so different in each
case that one is tempted tO think there must be something else at work,
and that the something else is a behavioral convention, a stereot ype . This
does not mean tO say, of course (imagination being the powerful force it
is ), that the beha vior is insincere. Let us go back to the ndop example, for
instance. If what I am tempted t0 think happens does in fact happen, then
neither the numbing effect of the music, nor that of the dance , supposing
the y exist, is sufficient to explain the collapse - any more than either suf -
fices on its own t0 explain entry into trance.
In tarantism, in which collapse occurs before as well as afte r the trance,
albeit in a form very different from that of ndbp, the role of the rh ythm is
t0 pro vide a regular (and rapid ) pulsation for dance music. As we know,
this dance music is regarded as a remed y for the crisis (and for the collapse
that was its manifest symprom ). But during the entire time that the taran-
tulee is entranced, the dance, in its turn, is intended to lead tO collapse (De
Martino 1966, 68 ) except at the end, when he is granted "grace " (ibid.,
73 ) and returns t0 his normal state. De Martino shows how important are
the intensity and proximity of the sound: "During the upright phase,
whenever the tarantulee - still dancing tO the beat of the music - lingered
in front of the musicians , or amongst them, the drummer would hammer
out loud rhythmic strokes close t0 his ears, but still keeping time ... "
(ibid., 70 ). Bur in his general interpretation of the part played by music in
tarantism (ibid., 145 - 46 ), he lays the most stress on the moto perpetuo
aspect of the tarantella. After giving an account of an episode that "pro -
vides an excellent demonstration of the relation between the crisis and its
90 CHAPTER TWO

musical resolution" he concludes : "Every thing that occurs seems to sug -

gest that a certain rh ythmical order ing of so und s unleashes char most ele-
mentar y sign of life which is mo vement, while, at the same rime, the disci-
pline of char rh ythm prevents the mo vement from sliding into pure
ps ychomotoric conv ulsions ... . T arantism begins , runs its course, and
ends - as ritu al and action - in the very elementar y perspective of a fresh
breakthrough int o historical existence by placing itself , pro tempor e, un -
der the protection of rhythmic repetition, as certain and as predictable as
the orbit of a planet. ... " In chis case, we are far away from the "feints"
and "breaks" chat Mecraux cells us are so propitiou s to loa crises .
Now let us pass on to anothe r aspect of the rel at ion s between rh ythm
and tr ance . So far we ha ve examined chem only wi th respect to a certain
number of specific cases. It is equall y possible to consider chem from a
much more general point of view, and chat is what I pro pose to do now,
very briefl y. T o make things as general as possible , let us take three musi-
cal regions: black Africa , Southeast Asia, and Indian America. And I
should make it clear, because some ma y ha ve doubt s in this respect, that
howeve r impres sio nistic or disputable it m ay seem, this particular divi-
sion -in my opinion at least- does correspond to a certain realit y. Ir will
be agreed th at the musical styles of these three regions all possess rh ythmi-
cal asp ects, specific to themse lves , char can be decisive factors in identif y-
ing them. There are certain unmistakable ch a racteristics char en able us,
w ithout the possibilit y of error, to attribute a particul ar way of drumming
to one specific region rather th an to another. There is a rh ythmics particu-
lar to Amerindians"' that could nor under any circumstances be black Afri -
can. I would be tempted to go further and say there is an incompatibility
between certain fundamental aspects of American Indian rh ythmics, on
the o ne hand, and black African rh ythmic s on the ocher. The fact that in
the present state of musicol ogy it is difficult ro say exaccly what it is, does
not affect the matter. In the same way there is a radical difference between
the rh ythmics of Vietnam, let us say, and chose of Amerindia and Africa .
This being so, in all three regions w e find the same relations between mu -
sic and tr an ce (possession tr anc e, as is usuall y the case in black Africa , or
shamanic trance , as is the case for Amerindia; the difference is nor impor-
tant ). If one p articular rh ythmic is thought to trigger trance in one region,
and in another qu ite a different rh ythmic performs the same function, the
reason must be th at an y rh ythmic , or any rh ythmic system - for an y rh yth-
mic is a system - can do the job as w ell as any ocher. Or else , w hich comes
to the same thing , no rhythmic system is specifically related to trance. This
totall y contradicts the assertion by D anielou , quoted at the beginning of
this chap ter , that certain rh ythms possess the particular virtue of being
able ro trigger trance in all regions of the world. If this we re the case, it
would mean that the action of rh ythm - and therefore of music -in trig -
 Music and Possession 91
gering trance is of a ph ysiological order. Everything points to the contrary:
the rel ations between rh ythm and trance operate at the level not of nat ur e
but of culture.
As a conclusion to these rem arks on rh ythm in possession mus ic, two
· obse r va tions should be m ade . Fir st, co ntrar y to what we found our abo ut
instrume nt at ion, w hich , as w e saw, reveal ed no p articularities, the rhyth-
mics of suc h music have two frequentl y recurrin g characteristics: on the
one h and bre ak s or abrupt changes of rh ythm , on the other an accele -
rando crescendo, w hich recurs so frequently o ne might view it as a univer -
sal of possession mu sic. Second, agai n contr ary to w hat w e found to be
the case for instrument ation , w hich we ex amined on ly in and of itself, or,
if one prefers, from the point of view of aco ustic s, rh ythmics mu st be con -
sidered from a double view p oi nt: in and of itself, that is, as a me ssag e re-
ceived by the per son in the trance, or preparing to enter it; as a dance per-
form ed by the entranced person; th at is , as a m odal ity of hi s or her
perception. For dancing to music (o r rather "d anci ng the music, " if one
may say so ) is in fact perce iving it in a particul ar way . For the dancer, it
means reenacting it on his own account. It means retr ansmitting the me s-
sage in the form of mo vement and nor merel y recei ving it. One is tempted
to say that it means "act ing " the music rather than simply undergoing it.
At any rate it certainly means substituti ng a totall y or partiall y passive re-
lationship to music w it h an overtly active one . In a later chapter, w e shall
once again observe th at as far as t he relation s between music and the sub-
ject in trance are concerned, it is important to distinguish be twee n those
that ar e ac ti ve and those that a re passive. We shall see, in fact , that de -
pending on whe ther o ne is dealing wi th shamanism or w ith po ssession,
either the active or the passive relations are involved. Thi s is therefore an
imp ort ant distincti on to make. As for d ance , it roo will ha ve a short
chapter devoted ro it later on .
Let us now pass ro facts about the melodic o rde r , or, more precisely, ro
facts abo ut the o rga ni zation of the relations between pitche s, in other
words, ro fa cts about mode, using the wo rd "mode" here in its mos t gener-
al sense. Are t here musical modes speci fically propitious to trance, or
more frequentl y linked w ith trance than o the r s? The anc ient Greeks and
Chinese, the civilizations of Indi a and of Islam, all the "high cultures" in
fact, have assoc iated mu sical scales w ith emotions, pa ssion s, the cosmic
o rde r , and ultim atel y, w ith man's m ora l and ph ysical health. These rela -
tions we re establis hed for the m ost part, it is true, at the level of learned
music , and consequentl y do nor necessarily or directl y involve popular
melodic syste ms. In a civilization in w hich there are gro und s fo r distin-
guishing between learned and popular mu sic, possession usuall y o ccurs at
the popular level. The question thus remains: Are there any connections
between musical scales and possession~
92 CHAPTER TWO

In Greek antiquity, Aristotle observes (Politics 8. 7. 1341 b ), the philos-

ophers classified melodies according to three categories: "et hica l" melo-
dies, linked with var ious ethoses or, let us say for simplicity's sake, various
states of so ul ; "practical" melodies, lin ked to certain occupations; and
"enthusiast ic" melodies, assoc ia ted, as their name ind ica te s, with posses-
sion . Elsewhere in Aristotle's Politics ( 1342b ) we re ad :
"All Dionysiac transport arid all analogous agitations find their
expression in the flute [read: aulos] rather than in any othe r instrument,
and these emotions receive a melodic accompaniment in the Phrygian
mode, which among all others is best suited to thF.m. The dirhyramb, for
example, is in everyone's opin ion a Phryg;an merer, and to prove it,
experts in the matter adduce a mass of examples, among others, the fact
that when Philoxenes attempted to compose a dithyramb, the Mysians,
in the Dorian mode he was unable to do so, and fell back, following the
natural course of thing s, into the only suitable mode, the Phrygian
mode.
We know that the ent ire Greek system of modes ma y be r ega rd ed as
resting, originally at least, upon a fundamental opposition between the
Dorian and the Phr ygia n modes. Since it used the Phr ygian mode exclu -
sively, "e nthusiastic " music must have displ ayed very specific characteris -
tics . But what were the y? In other wo rd s, in what way did the Phr ygia n
differ from the D orian? This is a complicated question , and one to w hich
we sha ll return at greater length in a late r chapter; suffice it to say th a t
Samuel Baud-Bo vy's recent findings ( 1977 ) lead us to believe th a t in it s
"primitive" fo rm the Dorian was a pentatonic mode wit h o ut sem iton es
and that che Phr ygian was, on the contrary, a diatonic m od e w ith sem i-
tones. Th e difference, from the viewpoint of expressi ve re sou rce s, is con-
siderable. Should Baud -Bo vy's views be confirmed, we cou ld be certain
that the cwo modes we re clearly dist inc t from one another. The fact re -
mains, however, chat Plato, who of all the Greek w riters tells us th e m os t
about the relations between music and possession, does not even refer to
the Phr yg ian mode when dealing with the subject. Just as he so frequently
cites the aulos as the instrument specific to mania, so he remains as utterl y
mute on the matter of modes. This fact is extremely important, and , com -
bined w ith a number of othe rs (to be treated at length later ), su ggests th at
Aristotle's views on the effects of t he Phrygian mode are but a much later
inte rprecac ion and rationalization of che rel at ions between music and en -
thus iasm. Placo, for his part, sticks to the aulos . Aristotle also acknow l-
edges that this instrument plays a great role in triggering trance (Politi cs
8 ). But the aulos is also the in strument of Marsyas, as Plato tells us in his
Symposium (215c ), and Mars yas is a Micr as ian. In a roun d abo ut way, w e
are back in Phr yg ia .
Music and Possession 93
So w e find th at in Hellenic Greece, not only the musical mode of posses-
sion but also its favored instrument were seen as coming from elsewhere.
Maurice Emmanuel ( 1914, 440 ), caking his lead from Aristotle, under-
scores chat Greek music is divided into two antithetical worlds, char of the
Dorian, "harmony of European Hellas " and that of the Phrygian, "music
of the Asians, imported by the autos pl ay ers of Phr ygia ." One easily recog-
nizes chat broad antitheses of this kind - and the one first made by Nie tz-
sche between Apollo and Dionysus naturally springs co mind - are rather
reductive; so let us not give this one any more importance than it actually
merits. However, we do not lend it coo much credit if we simply note
chat - as many ocher texts confirm - the music of enthusiasm or Dionysiac
mania was strongly felt co be Phr ygia n. It was indeed from Phr yg ia (or
Thrace, but in any case from As ia Mino r ) char the Dion ysus cult spread
(Jea nmaire 1951, 22 ) or, as the Greeks themselves believed, chat it or igi-
nated. We are therefore justified in saying chat although Dionysiac music
was Phrygian in instrumentation and mode, it was not because the musi -
cal characteristics of this instrumentation and mode in and of themselves
(chat is to say, their particular timbre or sec of inter vals ) were thought to
h ave any remarkable effect on triggering trance, bur rather because they
were the clearest sign of Dionysus' identity. Thi s would in any case co ncur
with the general log ic of possession. For what in face is possession other
than an invasion of the field of consciousness by the other, chat is, by
someone who has come from elsewhere? Insofar as he is the ocher, Dion y-
sus is at the same time an elsewhere; whether he is or not reall y is of litcle
importance. This is how he was thought to be, an d how his music was ex -
perienced. In Dahome y, people of Gun or Fon origin speak ago 46 whe n
the y are possessed by a vod un of ago origin . Elsewhere, people who or -
dinarily speak a given African language talk in Arabic if the spirit possess -
ing chem is thought co be of Arab origin. When she was possessed by Beel-
zeb ub, Je anne des Agnes spoke, it is said, in Hebrew (Certeau 1970, 66 -
67 ). What is true of speech could also be true of chat ocher language, mu -
sic.
Wha te ve r ou r interpret ation of the faces m ay be, it seems clear char
among che Greeks the music of Dionysi ac possession was characterized by
its use of one particular mode and no ocher. A comparable situation ma y
be found among the Thong a of southern Africa . Among chem, posses -
sions a re said co be caused by foreign spirits, either Zulu or dau. Songs
relating to possession by Zulu spirits are pentatonic , we are cold by T. F.
Johnston (1972, 13 ), whereas chose relating to possessions by Ndau spir -
its are heptatonic. The face chat one rathe r than the ocher scale is chosen is
directly related to the fact that Zulu music is pentatonic and 1dau music is
heptaconic. In other wor ds, the scale is not chosen for its particular me -
94 CHAPTER TWO

Jodie potential or expressive resources , but as a sign of the identi ty of the

spirit with which the melody is connected. This considerably strengthe ns
the above interpretation of similar facts among the ancient Greeks .
With respect ro tarantism, Diego Carpitella ( 1966, 361 ) observes that if
the m usicians are to be believed, "the tarantulees are incited to dance" by
certain tonalities only, "and remain insensiti ve ro others." This is because
each tarantula "likes its own tune." Consequently the repertoi re contains
"a traditionalized series of tunes and songs from which one must choose
the one (or ones ) best suited to each particular case" (De Manino 1966,
153 ). At least, th is was the situat ion some time ago, when tarantism still
flourished in southern Ital y. Nowadays, because the cult is in a sta te of
decadence, this "aspec t of tarantism is on the decline and has disappeared
from general awareness," Carpitella observes (1966, 351 ), and adds that
during his fieldwork, in 1959, his own observations did not confi rm what
the musicians had told him. Whatever the present state of things may be,
let us keep in mind th at, tradition ally, the different tunes played for the
tar antulee we re in principle related to the different species of tarantula
that may have bitten him or her, and the y thu s served to identify the par-
ticular one respon sible for the bite. Moreover, the tunes va ried in their key
signatures. Her e, then, we are confronted with evidence showing that mu -
sic is identified not with the region or people whence the spirit regarded as
respons ible for the possession o rigin ated, but \\'ith the species of the crea -
ture whose bite triggered the crisis. The choice of key or mode is thus de -
pendent in all three cases (the Greek s, the Th onga, and tarantism ) upon
considerations external to the music itself, and not upon the use of any ex-
pressive possibilities the music ma y possess.

Everyday Language
To conclude, let me turn ro another aspect of the matter, even though it
really concerns timbre rather than melody . Writing about the "low notes"
of the single -stringed fiddle, Rouch ( 1961, 141 ) tells us: "Accord ing to the
Songhay, it is these raucous, wavering, and tragic notes that speak to the
gods, that summon them and calm them. " A compa rative study of both
profane and sacred tunes th at make up the repertoire of the fiddle (godye )-
which can itself be used as either a sacred or a profane instrument - wo uld
let us formulate some judgment about the use of these notes in the "spirits'
tunes" and would thus help isolate the melodic system used in the Songh-
ays' possession music, if such a system exists. Similarly, among the Wolof,
one could make an inventory of the mottoes sung during the ndop and
then atte mpt to isolate the stylis tic characteristics of this repertoire.
Among the Yoruba, similar research could be applied to the mottoes that,
Music and Possession 95
in this case, are played on drums, rather than sung, or played on the fid-
dle. For my own part, however, I doubt whethe r such researc h- w hich
must still be done - would succeed in isolating styles that could be clearl y
distinguished from the style of the profane popular tunes belong ing to the
same literary genre. At the risk of jumping ahead too fast, I will say that ,
on the basis of the possession music I have heard so far (some of it from all
over t he world, but in particular that of the Songha y, the Wolof, the Gun,
and the Yoruba ) it does not seem at all probable that any such music uses a
musical system specific to it; that is to sa y, one chat constitutes a separate
and distinct system within the general musical system to which it belongs.
Indeed, paradoxical though it ma y seem, it is quite logic al chat possession
should speak, on the music a l level , the language of the ever yday. T he rea -
sons for this will be in vestigated later.
The same is not true , howe ver, of initiation and initiator y trance, which
are expressed , on the co ntr ary, by a specific mu .;ical lan guage character -
ized by its deliberatel y unusual nature. Ac lease , rhis is what we find both
in che Benin vodun cult and in the Brazilian cand om ble, as we saw earlier.
In Benin , the songs relating to initiator y trance are distinguished from all
the o cher music of the region by their ver y unusual use of chromaticism.
Are chere ocher such exa mples? One ma y w ell wo nder w hether a certain li-
turgical chant sung in the Buddhist repertoire, described as constituting an
"essentiall y ecstatic ""' music and performed in a scare that could be one of
trance, should not also be placed in the categor y of initi ator y music. This
chant has, of course, no connecti o n with chose alre ady described for
Benin , but it too makes systematic use of chromaticism. Compared to the
usu al Buddhist music one hears, it does strike one as being very unusua l
indeed. On the contr ary, the hymns for the genii sung during possession
seances in che Vietnamese hau bong cult, as descr ibed by Simo n and Si-
mon-Barouh (1 973 ), belong to the most banal popular st yle. ' 8 Whereas in
both Afr ica and Asia trance chat is not linked with possession exp resses it-
self in a musical language constituting a separate and d istinct system with -
in the population concerned (I hav e given two ex amples , but the list could
certainl y be extended, particularly in the area of shamanism and of com-
munion trance ), contrariwise, possession trance seems to require, every-
w here and at all times , a form of music belonging to the most ever yda y
and popular system. Despite appe arances, wh at I said a moment ago
a bo ut the use of Phr ygian music by the Greeks and Zulu or of dau music
by che Thong a does not vio late chis rule. Among the Greeks , the Phrygian
m ode, de spite its foreign origin , w as perfectl y inte gr ated into the general
musical system and was as much part of che popul ar tradition as the Dori -
an mode. And in a ll likelihood the situation is the same among the
Th o ng a . ' 9
96 CHAPTER TWO

WHAT DOES THI S MUSIC MEAN?

T he brief (and very incomplete ) survey I ha ve just provided of some of the

for m al characteristics of possession music thus tend s to indic ate th at , al-
though such music does play a pa rt in triggering and maint a ining the
tr ance state, it does not owe its effect tO the properties of its music al struc -
tur e, or if it does, it does so only tO a ver y sm all degree. Let us therefore
abandon this aspect of music , by wh ich I me an th e signif ying side of the
musical sign, and turn instead toward the signif ied one.
What does pos session music sign ify? What is its meani ng? Wha t are the
words of the songs saying ? Wha t do the dance rhythms exp ress? Or, rather,
tO what dances do those rhythms co rrespond and what do those dances
signify? We can say in a general sense that the vas t majorit y of such songs
concern divinities that ha ve come to be embodied, o r th at are expec ted t0
do so. The y can concern these divinities in several different ways. Some-
times they are addressed di rectly to them , summoning them or, on the
contrary, asking them tO go away; sometimes they desc rib e them , usually
in a flattering way, and often by referring to their genealogies or their mot-
toes, wh ich comes to the same thing.
Jane Belo, in her book on Bali, does not provide very much information
about the texts sung during tr ances caused by the gods. But th e songs fo r
the popular trances brough t abou t by sanghyang sp irits (the sanghyang
who inhabit the dolls and the little girl dancers on the one hand, and the
sangh yang of animals on the other [Belo 1960, 180 - 225 ]) fit in well wi th
what I have just said . Judging from Simon and Simon -Barouh's ch ap ter on
the subject, the same is tru e of the "hymns of in voca tion to the genii " sung
in Vietnam during ha'u bong ceremonies. In Tibet , the chant used to trig -
ger the oracle's trance is formed of two p arts ( 1ebesk y-Wojkow itz 19 56,
429 ). Th e first is an "invocation to the dharmapdla who is aske d to t ake
possession of the medium. T he invocation usuall y starts w it h a detailed
descript ion of the dharmapala and also of the abode in which he is sup -
posed to reside. " T hen follows "a litany ... p raising the dharmapala -
who by now should have occupied the body of the oracle - enumerating
and eulogizing his vario us capabilities" (ibid. ).
In Madagascar, the songs (antsa ) sung to summon th e diviniti es
(tsumba, a dialectical va ri ati on of tromba ) are "fairly slow." Th e wo rds
"aim to describe and flatter the god" (Koechlin, discography ). "T hese sol -
emn chants" alternate with other so ngs accompanied by a board-zither, an
instrument "also used on a schooner, to summon the w ind s when the boat
is becalmed." Once the tr ance is well established and the dances take on a
purel y playful character, it is dynamism of movement that is mainly re -
quired . The importance of the words decreases, and the y are often re-
duced to simple exclama tions. But for slower dances, the words ma y ac-
Music and Possession 97
quire importance again, and the singing consists of "success songs,"
occasionally humorous in content but no longer connected with posses -
sion as such. Among the Thonga of Mozambique, the importance of the
rh ythm and sound level provided by the drums in possession seances have
been strongly emphasized by H. A. Junod, as we have seen, but "that
which is the most essentially necess ary," he also writes "is the singing"
(1913, 441 ). Thi s is provided by experienced possessed persons: "The y
address the spirit in laudatory term s, trying to cajole it by flattery, to get
the right side of it, and thus to induce it to grant the signal favou r of a sur-
render " (ibid. ). Such phrases as "Rhinoceros, thou attackes t man!" show
that these songs obviously fall into the categor y of mottoes.
In Ethiopi a, "apart from the songs used to evoke the ziir, which are
called wadiigii like the gatherings during which the y are performed, " and
apart from "songs with prophetic intentions . .. often contributed by
professional possessees " and consisting most often of "a very general la-
ment about vicissitudes of the times," Leiris (1958, 43 ) also describes a
great man y songs "performe d for the purpose of entertainment." Else-
where, however, (ibid. , 80, nn. 2, 93) he m akes reference to a "song in-
tended to make the ziir descend onto a novice" and also mentions-
though w ithout expressly linking it to the singing - the use of mottoes
(jukkar ii) which he defines as a "fairly free accumul at ion of stereotypes re-
calling Homeric epithets."
In Niger, in the bori, the songs used to trigger trance (and on occasion
to calm it too, when the possession is a false one ) are essentially mottoes.
At least this is what emerges from the work of Jacqueline Monfouga -
Nicolas (1972, 109 - 59, 160, 201 , ere. ). Still in iger, amo ng the
Songha y the "musical mott oes" (that is to say, the ritu al texts ) which are
sung "usuallv with instrumental accompan iment" (Rouch 1960, 135 ), bur
which are also frequently played on the fiddle without being sung, pro vide
the very substance of their possession music. "E ach divinity has one or
several musical themes specific to him," writes Rouch, who then ass igns
these themes and rexes, the dance steps associated \,·ith them, and even
such objects as altars, ritual furniture, and costumes, to one and the same
general category of "more or less materialized siglls of the di vinities. "
In Senegal, the bak, a "theme-song or mono-song is one of the rab's at-
tributes . .. . Ir is with the aid of the bak, a short and more or less explic it
phrase, char the sp irits are summoned one by one during the ndop and the
samp" (Zempleni 1966, 304 ). In the warming -up phase, chanting of the
bak plays a decisive role. Ar the start of the session, after the drummers
have played the sum mons , the ndopkat' 0 form a procession, enter, stop in
front of the drums, and "consul t over the choice of the first bak ." Then,
when everything is re ady, the wome n o fficiants, in order to achie ve "a suit -
ab le degree of excitation," "summon those possessees who are particularly
98 CHAPTER TWO

easy to excite, by manipulating the appropr iate baks (ibid., 399 - 400 ). In
the course of th e extremely important session known as the "naming of
the rab," whose description I quoted earlier several rimes, various bak are
intoned in succession. Ar the end, "it is when a new bak is called out by the
assistants char A. G ... suddenly utters a long cr y, then announces: 'Wali
diaye Sene 1' This is rhe name the y had been waiting for.'' We shall return
lacer to rhe role of chanced mottoes in rhe process of identifyi ng a deity.
Among the Dog on ,5' in rhe course of rhe ceremonial enthronement of
one of the priests of rhe binu (mythical ancestors ), when the priests hear
rhe drum theme of their own binu, "they move into the center of rhe circle
and dance more quickly, or give themselves up to a ritual crisis char a col-
league then calms wi th an applica ti on of sacred bastard-mahogany bark"
(de Ganay 1942, 41 ). What is happening here is that "when the binu's
rh ythm is played chat binu's nyama [life force ] possesses the priest, who
then performs the movements spec ific to the theme being drummed out."
"For each binu there is a specific drum rh ythm (binu boy )," 52 G. Dieterlen
writes. When he hears the rhythm specific to his binu, rhe priest goes into
trance, because "at the mere bearing of the drum, the binu's soul is thought
to rise onto [his ] head ... and provoke the crisis." 53 Here again, then ,
among the Dogan, we are dealing with musical mottoes (which must not
be confused with verbal mottoes, tige) that trigger possession. But posses-
sion can equally well be induced , not by rhythms pla yed on a drum, but by
songs. At leastthis is implied in a rem ark m ade by S. de Ganay ( 1942, 43 ),
who relates the trance of a binu pri est "crawling on his belly to imitate the
ammo [wa ter spirits] who move in the deep holes in the earth" to a so ng
about the god Nommo . 5 ' But at the level of myth , the onset of the trance is
attributed neither to rhe singing, nor to rhe musical instruments, nor to
the dance, bur to rhe sound of speech. Ir was when he heard Nommo utter
rhe first word intended for the first men that Binu Seru, the first totemic
priest, "underwent the first fit, analogous to those char still make all
priests seized by inspiration 'tremble' today" (Calame -Gri aule 1965, 97 ).
Here again, then , in every case a sound bearing a meaning is seen as being
ar rhe origin of trance.
In Nigeria and Dahomey , among rhe Yoruba and also among rhe Fon or
the Gun, the oriki of the o risha and rhe mimlan of the vo dun respectivel y
occup y an important place in the invocation s chanced to summon the di-
vinitie s to manifest themsel ves, or to make them dance once the y have
been embodied (Verger 1957; Rouget 1965 ). These o riki and mimlan are
nothing other than mottoes, which ma y be spoken, sung, or drummed de-
pending on rhe circumstances. In Brazil, in rhe candomble, each deit y has
his o r her specific "canticles," which "constitute, together with the so no-
rous drum rhythms char accompany chem, so man y leitmotifs, to borrow
a Wagnerian term, intended to attract the orixa" (Bastide 1958, 19 ). "As
Music and Possession 99
these canticles unfold," Gisele Cossard w rites , "so the init iat es go into
trance" ( 19 70, 107 ). T he te xts she quo tes else whe re ( 196 7, 172 - 7 5 ) dem -
onstrate that we are once agai n dealing primaril y with mottoes .
The sung in vocations requesting the divinity ro co me a nd take posses-
sion of the adept combine mottoes, exhortations, pra yers, and, on occa -
sion, insults - a sign of the familiarit y th at reigns between men and gods.
Thus among the Vandau: "The y are not afraid ro insult the possessing
spirit, like the chief's daughter who is mocked for being a prostitute, or
like the a lm ost untr anslatable song of the Mandiki " (Junod 1934, 281 ). In
ne a rb y Zimbabwe, among the Tonga, A . Tr ace y (1970, 38 ) reports rhar
w hen people grow tired of waiting for the medium ro become possessed, a
cert ain m bira player specializes in sing ing a song so insulting ro rhe mhon-
doro (spirit ), who is usuall y add ressed in flattering terms, that sha me
forces him ro "come our" imm ediatel y. Whethe r the y are prayers, praises,
o r insults, these songs are add ressed ro the de ity, and this is the important
thin g ro remember. The y constitute communication with him .
When the y are specific ro a part icular deit y, melodies pl ay ed on an in-
strument ha ve the same func tion ·as su ng mott oes: ss they are call -sign s. In-
deed , these melodies often are mere in strumental versions of the sung
mottoes which are deprived of their text; but, when the y hear them, "men
and gods also hear the w ords that relate to them " (Rouch 1960, 135 - 36 ).
Whether linked ro a text or not , the se tunes are used for dancing, and the
result ing dance constitutes rhe motto 's choreographic aspect; it express es
the deit y's personalit y just the same, but in movement rather th an wor ds.
W hether it is a rh ythmic theme pl ayed on a drum o r a melodic theme
pl ay ed o n , say, a fiddle , rhis theme, when linked to a pa rticular deit y, has a
"signified " referring to rhar deit y, either bec ause iri s related ro him or her
by so me expre s ive correspondence (it depicts the deit y) or because it sim-
pl y has been arbitraril y attached to him o r her. Thu s the rh ythmic themes
specifi c ro the vari o us orisha in th e cando m ble are de scribed by G isele
Co ssa rd (1967, 177 ) as "dram atic, " "full o f viva cit y," "ag gressive, " "tem -
pe stu o us"- each o f these ad jecti ves corresp o ndir. g ro the nature of the di-
vinit y in question . On the other h and, I am un able ro sa y w hether or not
rhe "tunes o f the genii ," which constitute the repertoire of the hole y (genii )
cult amon g the Songh ay, also hav e an express ive co ntent . In rarant ism,
each variet y of tarantula has irs own particular tune. But is the co rre spo n-
dence in this case expre ssive o r is it purel y arb itr ary ? We do not know . We
do know, howe ver , th at in o rder ro cure rarantulee s "th e piper or zither
pla yer pla ys for them va rious motifs related to the qu alit y o f the venom, so
char the y are r av ished by the harmon y and fascinated by w hat the y are
he aring, and the venom either di ssol ves awa y within the bod y, or is slowl y
eliminated through the veins " (earl y sixteenth -century text quoted by De
M anin o 19 66, 144 ). Since each tarantula "likes its own tune" (ibid.,
100 CHAPTER TWO

15 3 ), there is "a traditionalized series of tunes and songs from which one
must choose the one (or ones ) best suited to each particular case." It is the
"r ight" music, in ocher words the one that best fits the case, that "will
make the tarantula scazzicare (leap about )" (ibid., 65 ), just like the appro-
priate cant icle in the candomble or bak in the nd op will attract the corre-
sponding orisha o r rab. But alth ough there are different va rieties of taran-
tell a , each differing from the other in rune, key, and rh ythm, it seems that
the dance - or rather th e sequence of the <lance's va rious movements-al-
ways remains the same, no matter which tarantula is involved. One of
these move ment s consists in imitating the spider. Ir is thus a figurative
dance, as are so man y ocher possession dances in Bali, in Viet nam , an d in
Africa. Depending on whether it invo lves a warrior (swo rd dance ), a lov-
ing wo man , a wild animal symbolizing power (tiger or panther ) or fertility
(snake ), the d ance w ill naturall y differ in character, and with it the music
that sust ains it. And in this case one is also dealing wi th more or less di-
rectl y figurative music .
The situation offered by rarant ism , in which the same dance is per-
formed to different tunes , corresp onds among the Mundang to an inverted
situation where, on the contrary, different d ances a re performed to the
sa me music. The film and rec ordings made by Adler, Pineau, and Zem-
pleni, already menti oned earlier , show that different possessed women all
dance at the same time, one the panther dance, one the sun dance, one the
rain dance , and still another the milling dance. In a small group, the y all
dance to the same music. This music va ries only in its details and, as we
have seen, nothing shows it is linked to one or the other choreographic fig-
ure. We are thu s dealing here w ith music that , in contrast with the figura-
ti ve music in rarantism , one might describe as abstract. Another example
of nonfiguraci ve music is the adarum rh ythm already mentioned several
times. The saint's daughters, who ordin ar ily respond onl y to the summons
of the rh ythm specific to their own particular orisha, all go into trance
when the adarum is pla yed. In the ndop, wh ile the rab's bak (the genius's
tune ) is what most distinguishes one genius from another (Zempleni 1966 ,
405 ), and while each class of rab has its own mimicr y (its "identificator y
beha vior "), contra riwise certa in d ance mo vements, p articularl y those of
the "labor of cri sis, " together with the drum mu sic directing them , of
course, are the same for all rab. This is true among the Songha y as well:
"Although each H oley ha s one or several specific mottoes, one or two spe-
cial musical themes, the d ances chat lead to his incarnation are not particu-
lar to him " (Rouch 1960, 146-47 ). The se dances comprise several distinct
phases , characterized by different figures and following one another at a
steadil y increasing pace until the y finall y lead to the crisis , but here again,
as with the "labor of crisis" in the ndop, the drumbe ats, and not the pl ayed
o r sung melodies , control the dancing. And since the same drum music
Music and Possession 101
wo rks for all the hole y or for all the rab and is thus impersonal in a way, it
does not function as a music al motto, 56 which is also the case with the
adarum rh ythm. The list of such examples could be extended consider-
ably. It is consequently neces sa r y, wi th respect to musi c accompan ying
possession d ances, to distinguish be tween sung or pla yed themes on the
one hand, wh ich are u su ally musi ca l mottoes, and on the othe r drum mu-
sic which m ay or may not function as a m ott o. This does not, howe ver , al -
ter th e fact that in practice the music may be an intimate mixture of melod-
ic music and of rh ythmic music, the two being associated in combinations
that va ry from cult to cult, and in whic h one or th e other so metime s pre-
dom in ates.
Let u s no w return to the musical mono, wh ich, as w e have seen, pl ays a
central role in possession. It can be defined as a sign w hose "s ignified " is
the god to which it refer s and whose "sig nifier" has three face ts: linguistic ,
musical, and cho re ographic. Th e signify ing power of rhis sign is peculiarly
extensive, since it involves sp irit and bo dy, intelligence and sensibilit y, and
the faculties of ideation and m ove ment , all at th e sa me rim e. This is evi-
dentl y what m akes it, fo r the adept , the most p owerful means avai lable
for identifying himself with the d ivinity possessing him. In certain cases
when the deity 's unknown identit y must still be established , it w ill also be
the princ ip al means of ide ntif ying him or her.
Tar antism pro vides a particularly str ik ing example of this use of music
or, more precisely, of the musical motto, as a means of identifying the di-
vinit y (here, the spider ) responsible for th e possession. In rhe late nine-
teenth century, one could still find "i n aples twelve different tarantella
themes used to diagnose taranrulees in order to establi sh which one corre -
sponded w it h which particular case , and consequentl y trigge red the
dance," writes E. D e Martino (1966, 160 ), who elsewhere quotes a pas-
sage from G . Bag livi, the celebrated iatromecha ni sr of the late sevent eenth
century, describing "mu sical exploration" (or diagnos is) as follows: "The
musicians who were summoned asked the p atient the color and size of her
tarantula so they could adapt their music accordingly; but rhe patient re-
plied she did not know if she had been bitt en by a tarantula or by a sco rpi-
on. They began trying o ut their theme s: at the fourth, the tarantu lee im-
mediately began to sigh, and at last, no longer able to resist the call of the
d ance, she leapt half naked from her bed, w ithout a thought for co nven -
tions, and for three days kept up a sprightl y dance , after which she was
cured ." We find rhe .very same process ar work in the case of the Sardi n ian
argia, a va riant of tar ant ism that departs considerably from the Apulian
model. Like the t a rantul a in It aly or Spain, or the scorp ion , in Sardinia a
m ythical creature , rhe a,gia, is responsible fo r the patient's poisoning.
Thi s creature "is categorized under three dist inct species, rhe nubile, th e
wife, an d the widow, and th e treatment of rhe poisoned person differs ac -
102 CHAPTER TWO

cording to the type of argia that bit him or her. In particular, the widow ar-
gia, symbolically associated w ith the color black, always requ ires dir-
ges ... the musical exploration [is] intended to determine the type
involved in the given case (and ] it unfolds through the successive use of
musico-choreographic themes and songs traditio ·nall y assoc iated w ith
each type" (ibid., 214 -1 5 ). 57 The success of this musical "ex ploration" de-
pended on the skill of the musicians, as did rhe efficacy of the cure, once
the "right music" had been found, De Martino also rel ls us (ibid ., 161 ).
This "right music" is, in fact, to give another example, what Plato
speaks of when he describes rhe Corybantes "fall ing into a trance when
they hear the melody specific to the divinit y by whom the y are possessed"
(Jeanma ire 1951, 134 ). And again, it is the "right song," played for the
medium among the Tonga that pro ves irresistible: as soon as the deit y
hears it "he is unable to refuse to incorpor ate him" (Tracey 1970, 38 ).
Among the Thon ga,' 8 as among the Wolof, it does not suffice for the
possessed woma n to declare the name of the divinity possessing her - a
crucial episode in her initi ation , as we saw earlier: she must also sing the
deit y's theme. In the ndop, after they h ave extracted the n ame of her rab
from her, the officiants continue to press the sick woman: "The y begin
shaking her again in order to get the rab's song -motto out of her. It was
some rime before the young woman began to sing in a staccato voice, the
bak required of her, and which all those present then repeated in chorus"
(Zempleni 1966 , 370 ). Among the Thon ga (or Tsong a ), things unfold
slightly differentl y. The patient sings his song , but he has to invent it him-
self: "Every possessed person invents a song which will be henceforth his,
and by means of which crises , or trances w ill be provoked or cured"
(Junod 1913 , 445 ). If the spirit responsible for the trance is a Zulu one, as
is most frequentl y the case among the Thong a, the words will be in Zulu,
even if the man involved does not kno w this language (ibid. ). This is an
extreme case of personalization of the sung theme , which, among the
Tonga, corresponds to rhe inverted case of "so ngs, wh ich while identified
with mhondoro spiri ts in general, are not tied to any one" (Tr acey 1970,
38 ). T his recalls what I said above about the adarum rhythm in the can -
domble, at the sound of which all the orisha, without distinction, go into
tr ance.

WHO M AKES THE M USIC A DIN WH .-\T STATE IS HE '

Who makes this music pla yed at va rious points in the ritual, which triggers
adepts ' trances and which makes them dance when they are possessed?
Who plays these instruments, who beats these drums, who sings these
so ngs? \'X'hatever the cult in question, in most cases it is important to make
a distinction bet we en two categories of individuals: those whose act ivity is
expressl y and exclusi vely to make music, in other words, those appointed
Music and Possession 103
ro m ak e it and whom I shall call the "musicians," and those whose activity
is ro make music only episodically, or ac cessoril y, or secondar ily, an d
whom I shall call "musicants" (musiquants ).
A sho rt parenthesis about this last word will perhap s be useful. In using
the French word musiquant, I was returning ro the verb musiquer, w hich
is hardly used nowadays, but which was used not so very long ago and fig-
ures in the well -known French Littr e dictionary, with reference ro excel -
lent sources such as Rousseau and Diderot. Present usage in the French
language does not m ake it possible to express in a single word the act of
making music (fa ire de la musique ). The same holds true for English. This
situation is often, and particul a rly here, a constraint . Using in French the
verb musiquer or in English the coined expression to musicate - which I
would propose as its equi valent - opens up a number of possibilities, such
as making a distinction between "make music" and "musica te," which is
not necessaril y the same thing , or between "musician" a nd "musicant," as
I have just done, or again, as I will do later, between a "musicant" and a
"musicated " person. Let us add that the series "ro music ate, musicant, mu -
sicated" could as well take the form of "ro musici ze, musicizer, musi -
cized ." The reader will choose which one of the two seems preferable . ..
o r less horrible to him.
E ver y, or almost every possession cult has two aspec ts , that of its pri-
\·ate rituals, re served for the initiates or those being initiated, and th at of
its public rites , in which both adepts and followers of the cult take part
and \,·hich more or less always take the form of a performance in which
possession dances constitute the central element. Fc.r simplicity's sake, we
will only examine this second aspect. One part of the music for these per -
formances - dance music , to be sure , but also the mu sic of overtures, pre -
ludes , interludes, postludes, di ·ertissements, and others - is furnished by
people especially appointed for the task and whom we shall call - whe ther
the y are instrumentalists or singers or both - the music ians . The y often
are professionals , and in this case the y a re always paid for their services.
The rest of the music - in vocations to the divinities, sung mottoes,5 ° calls,
pl ay ing of various accompanying or punctuating instruments, handclap -
ping, and so on - is pro\·ided by the adep ts (we shall later see wh ich o nes )
a nd by the spectators, all of \,·horn we shall call musicants when they ful -
fill thi s function. The difference between musicians and musicants does
not reside solely in that they participate in the general musical perfor -
mance in diffe rent ways, even though this difference is essential and is suf -
ficient in a nd of itself. The y also differ in their relation ro the cult and par -
ticularl y to trance. It is this last point that is of interest to us.
As far as the musicians are concerned things are relatively simple: they
do not, in principle, go into trance. Indeed , to do so would be incompati-
ble with their function, which is to prov ide for hours on end and some -
times on several consecutive days, music whose execution must continu -
 104 CHAPTER TWO

ously adapt itself to th e circu mstances . It is t herefo re im po rt ant th at t hey

should be constantly available and at the service of the r itual. T his is prob - .
ably the reason why these mus icians frequently are not adepts th ems elves.
Since they have never been possessed there is no fear that th ey will ente r
into trance. Paradoxically, then, these musicians, who seem to be the very
pillars of possession seances and without whom possession dance would
be inconcei vable, are in a way external to the cult. In any case, they occu -
py a place apart. Like the officiants, a nd in collaboration with chem, they
control the adepts' possessions and pla y a part in inducing chem , particu -
larly in the case of neophytes, bur , unl ike the officiants, they have neve r
lived through the experience themselves. In certain cults, however, it can
happen chat the y are adepts themselves. In this case the y have to be very
experienced in order to be able to resist possession in all circumstances.
Incidentall y, it does happen rhar they cannot resist ir.
Bue iris equally possible that they may ha ve every outward appearance
of being in a trance, when in face the y are not ar all. Speaking in general
terms about Haitian tambouriers (drummers ), Metraux writes: "Looking
at them, with turned -up eyes and contorted faces, hearing the rattle escap -
ing from their throats, one would th ink the y were in a trance, bur this ap -
parent frenz y is not due to any god. Ir is in fact rare for them to be pos -
sessed by a loa" (1958, 159 ). The musicians can certainly reach sta tes of
very great overexcitation, bur chis should nor lead us to confuse inspira -
tion with trance. However , as an example of a borderline case in which
the musician is not possessed but "inspired," let me quote w hat Jean
Rouch cells us when writing about the Songha y : "It is the fiddle r's (godye )
left hand that is 'inspired' (guided ) by the genii who are collect ively sum -
moned at the start of the ceremony by the 'hunters' theme' (gawey-
gawey )" ( 1973 , 532 ).
In Vietnam, in haubong ceremonies, the rask of providing the music is
assigned to a group of musicians, some instrumentalists, some singers,
\,·ho in principle are professionals (Simon and Simon -Barouh 1973, 40 ).
Bue, the authors of the description tell us, in the ceremonies they have ob -
erved in France , the musicians were "a mateurs," either adepts of the cult
o r simply '·disciples of the Genii." These musicians do not go into trance.
Th e instrumental music dominates the ceremony from beginning to end,
with preludes, interludes, and postludes framing the tunes to which the
genii dance. Contrariwise, the singing of the hymns breaks off during the
periods when possession is at its clima:v . It is \,·orrh noting, in passing,
that this seems to be a general rule: in cult s in which rhe music for posses -
sion dance is both instrumental and vocal, the instrumental music always
pre va ils and is always more continuous th an the voca l.
With respect to rarantism , rhe instrument al music, which predominates
ro a great extent, is pro vided by a small group of professional musicians

..___
Music and Possession 105
who specialize in this function . Quoti ng Kirche r , E. D e Ma rt ino w rites
a bout the "mus ician -t herapists ... who, in seventee nth -centur y T ar -
ento, w er e civil servants receiv ing set wages and w ho, in 187 6 still , w ere
regarded as persons of great so cial p restige" (1966, 21 2 ).,C ivil ser va nts?
Th is is pe rh ap s an exaggeration on Kir che r's p art , but they we re hir ed for
a p r ice and were the cause of great expense fo r t he patient or the pa ti ent' s
family. Unless I am greatly mistaken, it has nev er been reco rded tha t o ne
of t hese musicians was, or could be, a ta r antulee himself.
In Bali, when trances are caused by gods, the m usic ians of the village o r-
chestra are the ones who provide the m usic. In the case of tr ances ca used
by spirits, the music is mostly, if not exclusively, vocal. T he people of t he
village are the ones who provide it, and anyone who wants ro is free ro jo in
in the singing that regulates the possess ee's dance or pa nromime (J. Belo
1969, 40,202 ).
In the ndop, the musicians are "g riots," in ot he r wo rds professio n als
who belong to a particula r caste. T hey only play drums and do nor t ake
part in the chants intoned by the off iciant s and repeated in chorus by th e
ordinary adepts. The spectators participate by clapping their hands. In the
holey cult, among the Songhay, the princ ipal instrumentalist, who pla ys
the single -stringed fiddle, and the gou rd -drummer (s ) who accompany
him, are like\,·ise professional musicians. The fiddler - who is not a grior
( Rouch 196 0, 141, 144 )- is often also a zima or a priest of the cult. H e
plays and sings at the same rime . The situation is more or less compa r able
in the bori, except for the fact that the fiddler is nor a priest of the cult . In
Dahomey, in both vodun and orisha cults - and in many ocher cults in
black Africa as well - the m usicians, who are drummers, are not profes -
sionals; at the most they are semiprofessionals . T his is also true in the case
of the candomble . Bue whether in Africa or Brazil. ,,·herher profess ional
or not, whether they play the fiddle, the xylophone . or the drum, whe the r
they are pureh· instrument alis ts or whether the\' combine p lay ing with
sing ing, none of these musicians goes into trance.
If \,·e turn now · to the musicants, in other words the adepts 60 (officiants
o r ordinary member s ) insofar as they are musical actors, and their relation
ro tr ance when the\· are in chi role, their case is much more complex than
chat of the mu sicians . In a general way ,,·e find that the role of the adept as
musicant varies according co his seniority in the cult and his place in the
hierarchY . The more ne ophy te he is the less he rakes pan in the music; the
more he grows in seniorirv and import an ce the more he rakes pan in it, or
at least the more he is likely and able to rake pan in it , for the situation
changes from one cult ro another.
Ir would seem, moreover, that more often than nor (but not always, as
we shall see ), whether he is a neophyte or not , the ordinary adept does no r
take part, o r rakes very little part, in the music when he is possessed. For
106 CHAPTER T WO

him, the only expression of his possession state is dance and mimicr y. Th e
officiant, on the contrary, a veteran adept who is used to possession, can
sing and can play the ritual musical instrument he or she is carrying (if he
is carrying one ) while possessed. All of them, however, neophyte, o rdi -
nar y adept, and officiant alike, enter into trances at the sou nd of music
(instrumental, vo cal, or both ) made by the othe rs, and not by themselves.
This at least seems to be the rule, whatever the cult involved.
In the ndop, as in the candomble, the officiants sing the themes of the
deities while ringing bells, and their vocal and instrumental contributions
pla y an important role in triggering trance in the adepts in their charge.
But, and let me stress this, while fulfilling their functions as officiants and
musicants, which go together, they are not themselves in trance. T he ordi -
nar y adepts, for their part , also sing the divinities' tunes, but without play-
ing any instrument. In the candomble (Gisele Cossard, personal commu-
nication ) an adept is only permitted to "st art a canticle" after seven yea rs
of seniorit y. Before that she is only allowed co sing the refrain in chorus
wi th her companions . During the entire period of their apprenticeship,
which lasts seven years and ends with a special feast, new adepts thus par -
ticip ate in the music onl y passively, which is a very good illustration of
what I just stared on rhe relation of neoph ytes to music in general.
Although in most cases officiants only "musicare" during ceremonies
wh en the y are nor in a stare of possession , it can ne vertheless happen that
the y are in fact possessed at the rime. T he orisha cult in Dahomey provides
us with two ex amples of this exception , the first invol ving ins trumental
music, rhe second vocal. Am ong rhe Nago -Yoruba of the Sakere regi on, in
rhe course of the weekly ceremony for Ogun, god o f iron, 61 the god 's priest
goes inco tr a nce together wi th ocher priests ,,·h o are embodying kindred
deities. The music is provided by a number of mu sicians divided int o rwo
groups , drummers and bellringers on o ne side , a chorus of wo men on the
ocher. Th e officiants, four or five in number depending on the circum-
stances, all ho ld a pair of large iron bells \\·irh internal clapper s . During
the entire tim e the y are \\'airing for their tr an ce to begin, char is, while the
mu sicians (instrumentalists and singers ) summon the gods to incorporate
themsel ves, the off ici ants remain seated and silent . At the moment they en -
ter into trance, the y utter a great cr y and leap into the air, after which the y
emb ark all together upon a long ritual deambulation raking the form of a
quadrille that is half walked and half danced and during whic h the y con-
tinu ally bang their bells together in rime with their seeps, so char the result-
ing din mingles \\·ith the rest of the music. Here, then, we are dealing with
officiants in a st ate of possession who are ''musicat in g," but it should be
noted the y are not "musicating " in order to induce either other adepts or
themselves into the trance state. Th e same obser vation also applies to a
pr iest of Shango, god of lightning, who sings while he is possessed during a
Music and Possession 107
2
ceremony at which the first ya ms a re off ered to this deity. • Ir is only once
he is actuall y in tr a nce, wh ich was induced by su n g a nd drummed calls
pla yed by othe rs on his behalf, that he begins to sin g, limiting himself to
intoning the first wo rds of a song known to all and that everyo ne rakes
ov er in chorus.
Among th e Ga of Ghana (Kilson, 1971 ) the ade pt s of the tr aditi o nal re-
ligion (kpele ), which is a possession cul t , frequently sing kpele !ala (kpe le
songs ) during ceremonies when they are in a stare of possession. Bur , w e
should note, the y may also sin g them du rin g the same ce rem o nies w ithout
being in a trance (ibid., 26-2 7), and the se same so ngs can, m o reover , be
sung o utside of any ceremonies by t he la ymen. Whe n an adept wish es to
dance during th e kpele dances , which are in fact public possessio n se-
ances, she stations herself in front of the drummers and sings a kpele song;
th is serves as a sort ..,f prelude to he r dance and also gives the musicians the
required r hy thm. Whe ther the adept is entranced o r no r does nor affect the
way in wh ich thi s is done. T he autho r tel ls us no more on the subject;
what she does say suffic es to make it pla in, a n d aga in this is w hat we need
to remember, that the action of sin ging is unrelated a t this poi nt to the
triggering of possession. 63
What h as just been said concerning the Ga of Ghana could as we ll fit the
Gun of Porro -Novo (Benin ), a mong w hom I have often seen a possessed
vodunsi (adept of the vodun ) come and plant her self in front of the drum-
mers in o rder to sing the first ,,·ords of a given air, thus showing the musi-
cians the rune she wanted for her dance . Bur h ere aga in the singing had
nothing to do with inducing trance. W here as the m ajo rit y of these rune s
can be sung by the vodunsi w hether o r not in trance, some of them can
only be sung when the ad ept is possessed a nd w hen the dance rak es place
inside the convent. The study of thi s las t repertoire, ,,·hich is, of course,
much less acc essible th a n the first o ne, remains to be done.
On the ot her side of Afric a, in Ethi o pi a, the evidence provided by
Leiris, a lre ady mention ed earlier, co ncerning "so ng s with prop het ic inten-
tions . .. proffered by profess ional possessed pers ons," or "so ngs w ith a
moralizing tendenc y" su ng by Malkam Ayyahu whil e she is possess ed
(1 958 , 43, 66 ), again does nor im·alidate the rule th at t he possessed per -
son is not the one who sin gs in o rder to prepare his o r her entry int o
tr ance . The se songs, which a re not pos session songs, are nor sung by the
possessee during the ph ase preceding his trance, bur only af ter he h as al-
re ady entered it : "Malkam Ayyahu cove re d her he ad with a handke rchi ef
and, hav ing become Abba Yosef, sa ng seve r al mor al izin g songs and
quoted so me pro verbs," w rites Leiris. Describing entry into tr ance a nd the
preparatory ph ase preced ing it, he notes elsewhere (ibid., 83 ) that when
the pa tient has had hi s soul a lre ady "pa rtial ly 'veiled' by listening to the
wadaga song," 6 ' he rem a in s silent, so metimes for a long w hile, "w ith out
108 CHAPTER TWO

singing or clapping his hands w ith the rest of the assemby" before gradually
slipping into a stare of possession. Us in g his informant's own terms, Leiri s
adds in a note, to make things even clearer: "Befo re m ak ing them do the
gurri [the convulsive dance of th ose who are possessed], their zctr gives
them ams and makes them stay quiet during both the singing and the clap -
ping." One could hardl y express more clearl y the fact th at in the zcir the
possessed person is neither the musician nor the musicanr of his ow n
trance.
In Tuni sia, Sophie Ferchiou (1972 ) describes an extremely cu rious situ-
ation: that of women entering possession trances upon hearing chants that
are nor only not int ended for their ears bur also are being performed by
men who are trying to induce a qui te different type of trance in them-
selves . Her e is a brief summary of what l-appens . With in the context of
what in Maghreb is generally called the Saints Cult, men of a certain Sufi
brothe rh ood (that of the Saduliyya ) gather once a vveek in a sanc tu ary in
order to rake part in the well -known practice of dhikr . 65 Wh ile the se
adepts, secluded in a "totally dark" room, dance in place and rhythmicall y
chant the name of Allah, until they atta in the trance state, in a contiguous
room, women, who ar e "kept a t a distance" a nd who do "not participate
in rh e ritual," devote themsel ves to their own possession p ractices:

They do not chant with the men; they are interested, not in the words
but the rhythm to which they are chanted. When the rhythm becomes
fast enough they begin to dance, and, keeping time with the rise and fall
of the collective, guttural drone (made by the men ), they swing their
torsos forward and backward, back\,·ard and forward . Now and then
they make gestures that are thought to be dictated by the spirits
possessing them. . . Once the dhikr session is o,·er, when the men fall
silent next door, all the possessed women swoon to the floor. Attempts
are then made to revive them by murmuring in their ears phrases
intended to appease the spirits possessing them. [Ferchiou 1972, 56] .

The relation between the tr anc e of these possessees and music thus con -
forms to the usual scheme in the sense that rhe v a re nor the ones who make
rhe music. 6 6 T he unusual aspec t in this case is that although rhe music is
nor made by them, it is also not m ade/orrhem . In a way, the y appropriate
ir for themsel ves, wh ich shows th at man y combinations are possible even
within the boundaries of the restrictions established earlier.
The fact that the musicants, rhar is, the men practicing dhikr, are either
in trance or seek ing to enter it, does nor invalidate our rule, since they a re
not "music ating" for the wome n. On the other hand, the y a re, it is true,
acting as mu sicants for their very ow n entr y into trance. Bur in their case
the tr an ce is not a possession trance, an aspect to which we shall return in
a later chapter.
Music and Possession 109
Besides this example, w hich, with respect to the rel atio ns concerned, re-
presents the height of paradox, Sophie Ferchiou provides yet anot her ex -
ample that is so re assu ring in its classicism it deserves to be quoted. This
time she is dealing wi th possession dances performed in another sanctuary
and by another brotherhood, that of the Tidjaniya (ibid., 57ff. ). Here, the
music is provided by four women singers seated around a drum. These
wo men are "descendents of the saint" (ibid., 59 ), as are the two you ng
girls who rake care of the possessed women and make sure they do not
harm themselves. The musicians and the healers "never become possessed
themselves; they are immune on account of their birth, since their ances -
tor has tr ansmitted to them a little of the power he has over the spir its,"
Ferchiou tells us . These women musicians who never go into trance thus
confirm the rule observable in the majority of possession cults, since it
conforms to the inherent logic of possession itself.
Finally, let us look at the case of adepts who are still neophytes. Among
the Thonga of Mozambique, as among the Wolof of Senegal, we have
seen that the first ritualized crisis, a very dramatic event, is intended to
make the sick person, who is completely entranced, utter the name of the
spirit possessing him or her, or, in othe r wo rds, to identify that spirit.
After having uttered the spirit's name, the possessee must then sing his
tune. Among the Wolof this tune is a bak, that is to say a motto-song al-
ready known to everyone and long associated with that particular divin -
ity. Among the Th onga, on the contrary, the possessee must make up the
song. 67 Be that as it may, we should note that in both cases, even though
the neophyte in a state of possession does sing during the course of his or
her first ritualized crisis, the singing does not serve to trigger this crisis
but, on the contrary, to induce and express its resolution .
The same observation applies among the Lovedu of southern Africa,
neighbors of the Thonga, whe re the first ritualized trance also shares
many features in common with that of the Wolof . The main purpose of
the seance is to make the patient dance to the point of exhaustion in orde r
to force the spir it to manifest itself. Quoting Krige ( 1943 ), Luc de H eusch
writes the following: "The first exit of the pathogenic spirit is painful. The
patient sways when it approaches, his movements become more and more
frenzied, he drags his feet with great effort and finally collapses in a
trance. He then chants the song for the arrival of the spirit who is sa luted
by the attendance" (1971 , 265 - 66 ). Among the Kongo, quite some time
after the first ritualized trance, at the end of the initiation period, during
the coming-out feast which takes place at the end of three months, the
neophyte is invited by the master of ceremonies to sing in order to appease
the spirit responsible for her possession (de Heu sch 1971, 263, quoting
Van Wing 1938 ). Here again, in both cases, we see that the point is to re-
solve the crisis, not to induce it.
110 CHAPTER T\VO

So far we have seen adepts acting as mu sicants and not as mu sicians , in

the sense that singi ng or playing an instrument is only but one asp ect o f
their activity, and not its principal goal. Among the Mussey of Chad (Igo r
de Garine, see filmography ), on the other hand, the adepts do seem , on
certain occasions at least, to act as musicians properly speaki ng in that
making music appears to be their essential activity . When taking part in
certain rituals, the college of possessees (men and women ) constitutes a
veritable band of instrumentalists in which eve ryone plays the gourd -
horn, an instrument characteristic of possession. All of them are en-
tranced wh ile they play, and the y do nothing else. To my knowledge this is
the only case of its kind. The reason for this exception must certainly be
sought in the particular logic of the Mussey possession system, and in the
position this system occupies wi t hin the ir religious organiza t ion as a
whole . Bur whatever the reason for it, by behaving as musicians and not as
musicants the Mussey must be regarded as an exception .
What is it that should be remembered from the examination we have
just m ade of possession music in its relations with the people who make it
and with the state in which the y are at the time' First , that a large part of
the music is pro vided by people - the musicians - who do not go into
trance, who often are not themsel ves adepts, and who consequently are
marginal with respect to possession. Second, that possessees never figure
as either the musicians or the musicants for their own entr y into trance.
For the possessee, it is the music made by others that triggers, or helps to
trigger, his trance, ne ver the music he makes himself. This fact is signifi -
cant : in shamanism, as we shall see later, the reverse is the rule. Third, that
the adept's acti vity as a musicant increases only in proportion with the de -
gree to which he becomes more and more capable of controlling h is own
trance state. It is as though the adept becomes capable of "musicating"
only insofar as he has become less vu lne rable to the music. Th e ade pt th u s
passes from a stage of zero musical activity to one in which his mus ical ac-
tivit y is relati vely large. This coincides with his passage from the st atus of
neoph yte to that of officiant; in other wo rds, from a state of total sub m is-
sion to the deit y to one of relat ive independence. Last, that the offician t
"musicates" in order to trigger the trance of those adepts for whom he is re -
sponsible only w hen he is not in trance himself. When he is in trance, his
act ivity as musicant no longer serves his efficac y as an officiant, at least in -
sofar as he is in ch arge of others.
We thus find there is a kind of incompat ibility, total or partial depend -
ing o n the case, between the state of possessi o n and the act of making mu -
sic, or, if one prefers, between the state of possession and music as active
behavior, or, agai n , between being possessed and "m usicating ." When the
relation of the adept to music is examined at two particularly significant
moments - that is, when the adept is still a neoph yte and when, either as
Music and Possession 111
neoph yte or as officiant, he goes into trance - it turns out to be one of total
passivity. At the o ut set, it will be rem embered, we found th at the adept's
general relation to possession is, in a simil ar way, also fundamentally pas -
sive; it n ow turns out that possession is a form of behavior in volv ing a
double su bmi ssio n: to the will of the go ds, on the one hand, and to the ef-
fects of mu sic on the othe r. The inherent logic of the system basica lly re-
quires that th e possessed person be neither a musician nor a musicant, but
a "musica ted " person.

Uttering C ries
As a general rule, then , t he p ossessee is not a music ant. Thi s doe s not
me an, however, that he o r she is nec essa rily silent. "Crier comme un pos -
sede" (sho uti ng like a p ossessee ), fo r inst ance , is a current French expres -
sion . Utte rin g a cry often is a part of trance behavior. In tarant ism, for ex -
ample, the "crisis cr y" has been described as being someti me s "mo re like a
yelp than a human cr y" (De Martino 1966, 121 ), some time s, o n the con -
trary, as approa chin g a sort of music . According to D . Carpitella, to
whom we owe the ethnomusicological obser va tions appearing as an ap-
pendix in De Martino's wo rk , "the cha racteristic cry of the tarantulees [is]
based on two syllables 'A-hi ' in acco rd ance w ith a time -rat io that ma y be
codified as : A = hi! A - hi! A + hi! T his t ime- rat io de velops within a tes-
sirura going from approximately the inter val of a dimin ished seco nd to a
nint h" (ibid., 365 ). Others have referred co this cr y as a "ba ying" so und. In
t arantism then one has several ways of utterin g crie s when one is en-
tra nced .
The same can be said of the vodun and orisha cults in Benin. Frequently
the priest of Ogun emits, at more or less regular intervals, little yelping
sou nds. Adepts who are possessed by Sakpata, on the other hand, will at
certain moments utter modulated cries of extraordinar y violence th at very
closely resemble those just described for tarantism. This cry someti me s,
but not invariably, acco mpanies mimicr y quite clearly meant to represent
the efforts of parturition. In Niger, as Rouch 's films Jhow, it often happens
that those possessed by holey, and even more so those embod ying hauka,
do literall y bark when the y are in the mos t viol ent phase of their tr ance.
Other, similar examples could be cited. Moreo ver, we all know th e great
importance th at bioenergy adepts attach to the cry in their quest for liber-
ation through trance.
There is no doubt, therefore, that utte ring cr ies is a very signif icant
manifestation of trance, and that it needs to be decoded o n several levels.
For th e moment, let us simp ly state that in and of itself, the trance cry mer -
its a book -leng th stud y.
112 CHAPTER TWO

MUSICIA S Ai D THE P OSSESSED

What is the rel at ion of the possessee, n ot co the music this time, but co the
musicians? It is essentially characterized by the submission of the former
to the latter, and this follows log ically from w hat we have jus t seen.
In H aiti, Metraux tells us that "the tambourier [drummer] ma y not be a
professional musician, often is not even an initiate , but he nevertheless is
the m ainsprin g of the entire voo doo ceremon y ... a talented drummer
can indu ce possessions o r halt t hem just as he w ishe s." (195 8, 159 ). This
show s ho w great his power over the adepts m ay be!
Durin g ndbp seances, whe re , as we hav e seen, :he crisis and sub sequent
collapse are crucial, th e honor of "b rin ging down" mai nly falls on the first
dru mm er (Zempleni 1966, 416 ). When the possessee has alr eady been
d ancin g for a wh ile, and the cri sis is drawing ne ar , "h er whole bod y begins
co qui ver. At this sign the first griot mo ves ne arer and quickens the tem -
po." The "labor of crisis " th en ens ue s: "the pos sessed wo m an is at the mer-
cy of the drummers, w ho accompan y her move ment s and forc e them to
become ever m ore r apid and viol ent " (ibid., 415 ). In fac t , a close inter per -
sonal relationship develops at this point between drummer and posses-
see . 68 The drummer t akes ch arge of her, so to spe ak. Keep ing very close co
her, never leaving her side, concentrating on her slighte st mo vement s, in -
cess antl y obse rving her beh avio r in order to : speed up the tempo , or, o n
the contrar y, re lax it; select th e necess ary type s of beat; and adj u st the in-
tensit y of the stroke. Communicating th e rh ythm of the dance co her , he
holds the possessed woman in his sw ay and leads her int o the ever mo re
violent w hirlwind of hi s mu sic. But if he is able co lead her in this wa y, and
finally guide her where he w ishes, it is bec ause he h as been able to estab -
lish a close underst andi ng w ith her. It is bec ause he can follow her that he
is a ble co d o minate her an d impose hi s w ill up on her. He is the master of
the game, but w ithin a dialogue. H e speaks music and she rep lies dance.
Am ong th e Songha y, the principal instrumentalist , the fiddler , is also
so metime s a zima, a priest of t he holey cult . Jean Rouch (1 960, 14 8) simi-
larly emph as izes the personal rela tion established betw een the posse ssed
person and the musici an-priest du rin g the p reparat ion for and trig gerin g
of trance: "At an imperceptibl e sign, the prie sts and music ians sense th at
o ne of the summoned geni i is beginnin g co m an ifest itself ... the or che s-
tra now pla ys only one tune ... the dancers as a group contin ue to move
as befo re, but one of them has now become the sole objec t of the solic itude
demonstrated by the priests, \\·ho su r round him and rec ite motroes of
ever -incre asing efficacy . . .., Here agai n \\·e find th at in o rder to induce
tr ance in a particular pe rson the priests and mu sicians establish a spec ial
rela t ionship w ith him, "surro und " him, m ake him the o bject of their "so -
licitude," address themselves to him in an exclusive way, and become at
the same rime ve ry atten tive co what he himself is feeling.
Music and Possession 113
In his description of a memorable tarantism seance, at which he was
present with his team in Salemo in 1959, E. De Martino (1966, 69) dem -
onstrates the very close bonds woven between possessee and musicians.
The latter, "in a passionate musical offering . . . brought their instru -
ments to her ear. ... At one point during the phase c,n the ground, we
saw the rarantulee crawl over ro the feet of the viotinist and linger there as
though hypnotized; the violinist then knelt beside her, enveloping her w ith
his music to such a point that his bow seemed to be using the woman's
quivering body as its violin." In Bali, as we saw earlier, Jane Belo describes
how the singers move very close to the individual concerned in order to
make him enter into trance.
Instrumentalists or singers, professionals or not, when the y attempt to
trigger the trance sta te , the musicians, as we have seen, enter into very
close contact with the possessees. Communication is established between
chem , nor only at the level of the code involved (words of the songs, dance
movements ), bur also at the person al level, the emotional level of direct
person -to-person relationships.
Here again, however, there are distinctions to be made depending on
the type of cult involved . In ceremonies like the first coming-out of the
new initiates in a vodun cult, or the vodun's annual feast, or the "Cus -
toms" in Abomey (during which, as we shall see late r, princesses embody
the ro yal family's departed ancesto rs), the musicians and the vodunsi form
two separate groups, each of which remains compact, and from which no
one separates him - or herself at an y point. 6 9 The possessions in these cases
are, it must be remembered , of the kind we have termed "impersonal."
Similarly, the relations between musicians and possessees ope rate not at
the individual level, as in nd op, bo ri, or tarantism, in which possession is
a personal matter, but at the group level, a fact corresponding to the log ic
o f the system.
The power musicians ha ve over the possessees is not entirely due to
their talents; it also stems from the instruments the y pl ay, or rather, from
w hat those instruments represent . In the orisha cult, "the dru ms are very
much respected, for the y are nor just musical instruments. The y are seen
as being the voice of the gods themselves; through them one summo ns the
gods and, at the same time, replies to them" (Verger 1969 , 59 ). The y are
consecr ated, moreo ver, ju st as the adza bell of the cando m ble and the fid -
dle o f the holey cult are consecrated , to name but two examples. As a re-
sult of this co nsecr ation, Rouch tells u s, the use of this particular fiddle be-
co mes restricted "since even a single note pla yed upon it is a summons that
cannot go unnoticed in the invisible world " (1960, 144 ). This is because
the instrument is in fact equipped with a rattling device whose discreet
clicking "guides the music" to the genii's delicate ears . ' 0 In the bo ri, Jac-
q ueline Monfouga -Nicolas writes, the single -strin ged fiddle, which is "ca -
p able of forcing the gods to embod y themselves " (1972, 139 ), is "the most
114 CHAPTER TWO

sacred of all" the instruments (ibid., 159 ). According to information given

to Michel Leiris in 1931 (personal communication ), but wh ich needs to be
verified ," among the Bambara the water spirit (dyiro dyina ) responsible
for possession is supposed to come and position himself on an iron rod
placed in the water drum ; 2 (th e instrument used for possession ) after the
drummer has begun to play . The musical instrument is thus seen some-
times as the voice, sometimes as the more or less temporary seat of the di -
vini t y. Evidence of this sort is so plentiful and well known it is unnecessary
to cite any more . But the symbolics of musical instruments does nor inter-
est us here solel y as an element of the very complex system of signs consti-
tuted by a possession cult. Th is symbolic efficacy is twofold. On the one
hand it reinforces the effective powe r of communication these instruments
establish between not only gods and men, but also musicians and d anc ers.
On the other, it helps strengthen the power of the musicians who pl ay
them and thus increases their ascendanc y over the possessees for whom
the y pla y.

M USIC AN D D ANC E

Whereas, as we have just seen, the task of playing music during possession
ceremonies does not fall on the posses see but on the others, the possessees
are the ones who do the dancing. Dance should thus be seen as doubly
more important than music; first because it is the possessee's own busi -
ness , and second because music, or at least a large part of the music, is
pla yed for the purposes of dance. A theor y of the relations between dance
and possession is thus even more necessary than the one for w hich I am at-
tempting to gather together the basic elements in this book, an d w hich
will only concern music. It is hardl y worth sa ying that for all kinds of rea-
sons (chiefl y because dance is a vast subject that lies outside my domain ), I
will not even attempt to propose such a theor y here. n At the same time, I
cannot altogether avoid mentioning dance. But I shall do so only insofar
as it is indispensible to situating it in the perspec tive adop ted here.
In a number of cults it is possible to distinguish between two types of
dance. There are what I shall term abstract d ances, whose function is to
trigger trance, and figurati ve dances (or simpl y mimes ), whose function is
to manifest the possession state. Among the Songhay, the dance proper ,
which is made up of three "princ ip al mo vements " (Rouch 1960, 14 7- 48 ),
does not represent an ything - it is abstract - and remains the same rega rd-
less of which genius is expected to become embodied. This dance helps to
summon him: "the music, the words of the mottoes, the dance steps, trig-
ger the strange mechanism"; and, when a certain stage is reached, the
dance indicates that the genius is on the verge of becoming present . After
ha ving been grounded b y the fir m arking the deit y's descent, then "calme d
Music and Possession 115
by the attentions of the 'quiet women,"' the possessee ado pts the beh av ior
befitting the god he is embodying, and w h ich cons ists in stereo typ ed mim-
icry. In the ndop, besides the dance an d the "labor of crisis" result ing in
collapse and which are the same for all genii, there is also a mimicry rep er-
toire - or "identificator y behavior p atterns" (Zempleni 1966, 402ff .)-
having the same function as that among the Songhay. In theory, adepts of
long sta nding perform only the mimed dances, since convulsive dance, cr i-
sis, and collapse are restricted to neoph ytes .
In haubong ceremonies, the dancing also include s two phases (Simon
an d Simon -Barouh 19 73, 44 ): "The first is the descent of the Genii .... It
is reproduced identicall y in all cases; only the sign of the raised finger is
changed." While the drum "beats out the rh yt hm of possession," the wom -
en singers "summon the Genii" (ibid., 42 ). The dance consists of the
ade pt 's head "turning, slowly at first, then more quickly, drawing the
who le torso into its rotation. By me ans of a coded sign made with the
ha nd - the number of fingers raised indicating the order number ... the
Gen iu s .. . discloses his identit y." When the embodiments ha ve been
comple ted, the va riou s genii execute figurative dances: the sword dance,
flame dance, oar dance, unicorn an d lio n dance , and the "wa nds with pel -
let -bells" dance (ibid., 45 ). In the middle of this last d ance, that of the
Princes an d little Princes, "the Tenth Prince inserts an ep isode in which he
mimes a bowed musical instrument while the notes of the scale are sung as
an acco mpaniment. " Although in the haubong, contr ary to what we have
seen in the ndop o r among the Songha y, the d a nces contain no elements of
violence (the only d ance which does is that of a much -feared genius whose
descent is "heralded by violent somersaults [ibid., 57 )) ; there are clea rly
great similarities between these Vietnamese ceremonies and the African
ones just described. T he same could be said for tarantism, w ith its "regu -
lar choreographic cycle" ( Carpi tell a 1966, 359 ): the sword dance, spider
dance, co n vuls ive figures (that of the arched back in particular ), and the
seep "mo re properl y agonistic" consisting in rhythmic stamping. These are
just a few examples, but the y will situ ate the problem within the perspec -
tive of chis book.
It would no doubt be difficult - and possibly futile - to ~raw p recise
boundaries between abstract dance, figurative dance, and mimicry. A par-
ticular dance may seem to be nonfigurative only because we do not know
(or it is no longer known ) what it represents or symbolizes. These three
categor ies do nevertheless exist, and they furnish the elements of a system
of combinations that var ies from cult to cult, and which und oubtedly tells
us a great deal abo ut the underl ying logic of these cults.
For the adept, dance is the best pos sib le means of exhibiting, if one may
so put it , his state of possession, since his mo ve ment s, seeps, mimicr y, and
costume are in reality chose of the god inhabitin g him . But in certain cults
116 CHAPTER TWO

dance is also a means of compelling the gods to become embodied. "If the
music and the dance are pleasing to the spirits, to the point of ensnaring
their will, this is because they are themsel ves dancers who become carried
away by the supernatural power of the rh ythm," Metraux writes (1958,
168 ). "They danced, these powerful genii," Rouch tells us, "because the
music forced them to do so" (1960, 145 ). "A deity's supreme requirement
is that one among the faithful become his 'horse' by means of tr an ce , s·o
that he can come down among mankind and dance ," Gisele Cassard
writes (1970, 95 ). Music incites the gods to dance, but in order for them
to dance, the y muse have a visible support, and this forces them to become
embod ied . Whethe r we are dealing with voodoo, the holey cult, or can-
domble, the symbolic system is the same . Of course, this does not mean it
is the only one .
Whereas Rouch and Zempleni see dance as one of the principal means
(if not the principal means ) of inducin g crisis and fall (wh ich we must re-
member do not have the same function among the Songhay as among the
Wolof), Le iri s, on the contrar y ( 19 5 8, 18 ), says thatthe gur ri (the dance of
the ziir cult ) which is also characterized by violent mo vements, is the "sym-
bol of trance" rather than the "means of pro voki ng it. "" While referring to
it as "a classic procedure for inducing trance," Leiris obse rves that the
gurri has the particularit y of marking "both the departure and the arrival
of the spirit, " so that it is thereby assoc iated w ith both the beginning and
the end of trance. We must therefore rule out the ide a that its role is to trig-
ger trance in a mechanical way, as it were. Further on, Leiris notes that
"certain great zar ... do not make chose the y possess dance the gurri"
(ibid., 92 ), which is proof that entr y into trance can perfectly well take
place without dance . As with music, the problem is to find out whether or
not dancing pla ys a role in inducing trance. If we say it does, then we also
have to ask: Is this role real or symbolic? And in either case we still have to
as k whether this role should be attributed to the ethnographer's discourse
or to chat of the adepts themselves.
T he rel ations between the two aspects of dance -fi gurative and non-
figuracive - and the cwo forms of music-instrumental and vocal -
emerge so clearly in a note on the dyide (water spirits ) cult among the Ma -
linke of Mali pro vided by Cheron (19 31 ) th at a long passage of his text
merits quotation in its entiret y. Thi s passage also revea ls that, as in the
ndop, the female offic iant s (the "high -r anking members" ) want to avoid
the fall, which is to say, ultimately, the full trance state, since this is the
business of the "or dinary members," who a re, broadly speaki ng, new
adepts: a significant rule since it reflects an entire section of the cult's inter-
n al logic.
The dances are alwa vs accompanied by one or two dyidzmu (water
drums ) played by a wo m an, the dyidunu fo la (she \,·ho makes the water
Music and Possession 117
drum speak ); this music is supported by songs, whose tune and words
are given (donkili la) by a wo man musician, then repeated in chorus by
the aud ience. Each dancer executes steps and gestures, wh ich mainly
consist of rapid movements of the legs, arms, and head.
The ordinary members dance more often than the others; they enter
the circle when the woman singer intones the tune specific to their
particular genius, which immediatel y sends them into convulsions. The y
then perform their jinn's dance, a special dance thar simply mimics the
jinn's habitual actions; both song and dance are taught to the adep t by
the chief, the only person, given his infallible power of identification,
who knows the genius of the ordinary members.
As for the high -ranking members , they dance as little as possible, as a
precaution , for should th ey happen to fall to the ground during their
revels, this would be a sign of the imminent death of an adept. The y
cannot, howe ver, abstain from dancing altoge ther for fear that their
genii, thinking they are rebellious, would oblige them to do so. When it
is their turn to dance, they require the assistance of ordin ary members
whose dut y it is to watch over them and hold them up should the y fall.

If, in possession, dance oscillates be tween two poles, the figurative and
nonfigurative, the one being dance as identificaror y behavior, the other
dance as trance behavior, this is bec a use it pro vides the ade pt with the
me a ns of assuming his new personality and living intensely at the moror
level. Depending on the type of cult , one or the other of these aspects will
predominate, but both usuall y seem robe present; either simultaneously,
if the trance takes on both as peers at once ( as in the case of the spider
dance in tarantism ), or else al tern ately.
In possession, then, dance is a representation of the gods, in other
words theater-sacred theater, but also theater that one enacts not only
for oneself, as Leiris has shown, but also for othe rs. But dance also is
physic al exercise, as it is by definition, in any circumstance. Viewed from
either of these aspects it is esthetic activ it y and pl ay. But in all of its aspects
it is , above all, communication - w ith onesel f and \,·ith others.
For the possessed person it is a system of sign s both in relation ro him -
self, in that it "s ignifies " his borrowed per sonal ity for him, an d in relation
ro the group, in that it expresses this personality and marks ou t his role (in
the ordinary and th e theatrical sense o f this term ) in the "rep resentation"
of the world constituted by the possession ceremon y. ' 5 During the course
of the dynamic process presented by possession in it s succession of chang-
in g states, dance also "sign ifies " the stage that the possessee has reached in
his metamorphoses. Moreover, whereas it is by no means certain that pos -
session music has purel y ph ys iological effects on the possessee (or the per-
son about robe possessed ), dancing-be it in conjunction with possession
or not-undoubtedly brings about modifications in the da ncer 's state,
both a t the physiological and psychological level. These physical effects
118 CHAPTER TWO

are the ones soug ht after, and thi s explai ns why dancing can be seen as a
form of ascetic exercise and why it is often used as a trance technique . But
no m atte r how important its nature as a sign m ay be, or its symbolic func:
tion, esthetic po we r, or asceti c possibili ties, dance is still a motor act ivity
that finds an en d in itself. D anc e is always, at least in part and sometimes
despite appeara nce s, the ple as ure of dancing, of u sing one's bod y in play.
If it is nonfigurative, dance is pure physical expenditure. In this sense it is
already liberation, catharsis. If it is figurative, particularly when erotic o r
wa rli ke, dance is m anife stly a me ans of release; when it imitates animals,
it is also a means of rele ase, al beit through a symbolic ro ute. Here aga in it
is cathartic, this time in the classical sense of the term . At o nce "mi me sis
and catharsis" ( to borrow one of the subt itles of Jeanmaire's book ), dance
can ultimately be seen as the true realization of the state of enthusiasm .
The identification experienced by the individual with the god he embodies
takes place through dance and because of it. It is also throug h dance that
this identification is made manifest in the eyes of others. This second func -
tion is as important as the first, since apparently the very exis tence of pos -
session cults requires that possession be public behavior .
I spoke earlier, with reference to music, of the importance of the "right "
tune, which makes it possible both to identify the divinity and to make
him or her dance. If this tune is said to be "r ight," is it not so bec ause the
dance with which it is associa ted re veals, over and above the apparent be-
havior of the divinity, the deeper nature (o r the hidden as piration s) of the
possessed person? Or at least one aspect of hi.s or her n at ure . It seems, ac -
cording to Gisele Cassard, who ma de reference ro her personal experience
of the candornble (personal communic at io n ), that multiple possessions
would provide a means of ad ding light and shade co the inner portrait of
an individual by organizing, around the principal divinity he is embod y-
ing, a constellation of secondary di vinitie s (also identified by divination,
of course ) who might unveil oth er traits of his character. But the rel ati ons
of individual to divinity can be of ano th er kind, notably that of filiation,
as is generally the case with imperson al pos sess ions.
It can happen that possession trance is achieved w ithout music, and
consequently ,,·ithout dance, since the latter does not occ ur w ith out the
fo rmer. In an earlier chapter we already noted one exa mple of this for the
candombl e during initiation; but this happens much more frequentl y in
in spira tion o r mediumi stic trance th an in possession in the stri ct sense of
the word. This is conf irmed , among other examples, by the sit uation
found in southern Indi a, where the possessee's main function is ro deliver
oracles. T he possessed man, L. Dumont tells us, "is described as he who
'dances the god' ... " (1957, 350 ). But he adds that although "dance is a
means of inducing or re vealing possession, it is not indispens ibl e ro this
process. Certain kodangis [possessees] are thought to be able to provoke
Music and Possession 119
possession all by themsel ves. This is why they were sometimes consulted
wi th ouc either music or d ance , either during prelimin ary meetings, or on a
.day afte r a cult ceremon y for familial oracles."
Before puttin g an end to this r ather brief summ ary of the relations be -
tween dance and trance, shamanism must at least be mentioned. A musi-
cian (as w ill be shown in the next chapter ), the sha man is as much a dancer,
an d the role of dance in sham ani stic tranc e de serves in-depth trea tment.
Suffice it to say here that sha m an ic dance is as much of a show as is p oss es-
sion dance, although in quite a different m anner.
H oweve r paradoxical it ma y be, the only example I shall give will not
come fro m Siberia but from Africa, and w ill co ncern a case whose sha -
manic character remains to be proved. I refer here t0 the Z ande w itch doc -
to r who, w hen entranced, acts both as a di viner and a he ale r (seep. 24 ).
We owe to Eva ns-Pritchard (19 37, 149- 82) the most extraordinary de -
scription of a "seanc e" in w hich dance appears as a means of divination.
Indeed, for a Zande witc h doctor, divining is a highl y corporeal technique
(as Mauss would ha ve sa id ) ach ieved through dancing. "A wi tch doctor
does no t only di vine with his lips, but w ith his whole bod y. He dances the
quest ions which are put ro him," E va ns-Pritchard wri tes (ibid., 176 ). To
su m up ver y quickl y, he describe s th e witch doctor 's d ance ("violent, ec-
static . .. the most spirited performance of the art th at I have ever wit -
nessed" ) as being the princ ipal way of stirr in g and enhancing divinato ry
intuition within himself. Thi s is, to be sur e, an unexpected aspect of
dance, alt h ough it deserved to be cited.

MUSIC , GODS, ILLNESS

I hope the reader will forg ive me fo r now entering the re alm of co mm on -
places and basic truth s. At this particular stage in m y argu ment I feel it
necessary to recall a few of them. T hey will bea r first upon the manner in
which music is experienced, then upon the manne r in which the presence
of the gods and the weight of illness are expe rienced, since in matters of
possess ion the latter two frequentl y go roge th er.
How is music experienced 1 ' 6 We respond to it in several ways; to sim-
plify matters, let us say physiologically, psycholo;;ically, affect ively, an d
esthetically.
Ph ys iologicall y speaking (at the sensorial level ), alth ough music is
mainl y perceived by the ear, this is not the only path it can rake. Musical
vib r atio ns are wave mo vements whose ampl itude is relati vely la rge whe n
compared to the scale of the human body . Th e movement of th e obj ects
that give ri se to these vibrati o ns - or the movement that the y excite in ob -
jects, since the tr ansfer of energ y can take place in either direction - is al-
ways palpable and often even visible. It is thu s directl y perceptible as m a-
120 CHAPTER TWO

terial and concrete. A musical vibration can be something palpable. Hone

touches the soundboa rd of a viol in wh ile it is being pla yed, one can feel
th e sounds quivering against one's finger tip. If one nears one of the ex- ·
tremely large drums the Yoruba beat at their secret oro ceremonies, one
will hear the sounds through one 's abdomen-which vibr ates in sympa -
th y- as much as through one's ears. Similarl y, if in the same ceremonies,
one comes close to one of the small drums (whose tightly stretched skin is
wh ipped rather than beaten, prestissimo, with extremel y slender sticks ),
one's entire head vibrates. In an o rgan loft , when the organ plays loudl y,
one absorbs the music with one's whole body. T he whole world trembles,
the ver y air resounds. "To bathe in music " is not just a metaphor. It hap-
pens that we truly perceive it through the skin. Th e candle flames flicker-
ing in churches at the sound of the organ pro vided Louis Roger (1748) 77
with one of the observations on which he began constructing his theory of
the effects of music on the human bod y. The curious use that Tibetans
make of the vibr at ions of a drum skin 78 is also based on the same type of
observa tion. T o blow into a clarinet is ro feel the reed quiver between
one's lips. To ring a bell held in one's hand is robe in direct contact w ith
the vibr ations of the metal one is striking. To shake a rattle is to feel the
skin of the gourd vib rate as it is struck by the seeds inside. T o pla y a Jew 's
harp is to feel the met al rongue moving in one's open mouth.
But this external sensitiviq · is nor the only thing involved. One 's internal
sensitivity is also aroused by music and likewise functions as a path of re-
ception. It is well known that when we speak, and even more when we
sing, w e hear ourselves from inside. Bur in fact we are doing m ore than
hearing ou rsel ves sing. We are feeling ourselves sing. We are feeling our
larynx (let us say, more simply, our neck or throat ) vibrate, qui ver. And
this is true of many other are as of the body coo: the entire head, the thorax
and abdomen, the pelvic region. 79 Music is thus simul tan eousl y an anima -
tion of thing s and a palpitation of the being. Both are felt more intensel y
\\·hen one is making music than when one is simply listening to it. We
should be able ro say "ro act" music , as opposed to "un dergoing" it, for
these are indeed two very different ways of experiencing it. Be that as it
may, wha t we need to remember is th at music has a ph ysical impact upon
the listener and th at it produces a sensorial modification in his awareness
of being . This physical impact, of course, is what pop music is consciousl y
stri ving for. Through the din of va st amplifiers, such music obtains effects
of viole nce and acoustic turbulences never befo re ac hie ved. H aving told
us th at the amplification of pop music can "reach 10,000 watts" ( !), and
that it "was possible to hear the pop festival on the Isle of Wight from three
kilometers away," Alain Roux observes that "this amount of power acts
directl y upon the bod y and creates a feeling of participation that man y
people never att a in even during the sexual ac t. There is no resisting it ex -
Music and Possession 121
cept by flight. Amplified to this extent, the human voice affec t s the lar-
ynx .... The sounds of the electr ic bass (infra -sounds ) prod uce vibra-
tions localized in internal erogenous zones of the abdomen ... . T he
repetiti ve melodies and perpetual thrumming instantl y produce a lig ht
hypnosis" (1973, 130 ). Clearly, this is the kind of effect that certain forms
of possession music are aiming for; that of the ndop, for example. 80
Music is in essence mo vement. Its origins lie in bodily movements - to
sing is to mo ve one's lar ynx, to drum is to move one's arms, to play the fid -
dle is to move one's fingers along the stem and a bow across a string -and
in return music is an incitement to movement . Since by definition sound is
accua lized in the unfolding of time, its relations with itself are constantly
changing (even if it remains the same, for continuing is ine vita bly also
changing, since it implies change of duration ) and these changes are inte -
grated, at several levels, into the "th ickn ess·' of time. Even in its most im-
material aspec t- sound totally isolated from its source - music is per -
ceived as movement being realized in space. This is even more true when it
is made simulta neousl y with dance, or ro make people dance. To dance is
to inscribe music in space, and chis inscription is reali zed by means of a
constant modification of rhe relations between the va riou s parts of the
body. The dancer's awareness of his body is totally transformed by chis
process . Insofar as it is a spur to dancing , therefore, music does appear ca -
pable of profoundly modifying the rel ation of rhe self \,·ith itself, or, in
o cher words, the structure of consciousness .
Psychologically, music also modifies rhe experience of being, in space
and time simultaneously. Like the sound of speech, rhe sou nd of music de -
fines the space in which I am situated as a space inhabited by men, and at
the same rime it situates me within chis space in some particular manner.
Silence is rhe sign of an empty or motionless space - death or sleep. Sound
is rhe sign of a space that is filled and in mo vem ent. The sounds of nature
bring me information on the movements of n ature, the sounds produced
bv men bring me information abo ut the presence of men , and the types of
sound I hear cell me something about their activities : the y are cutting
down a tree in a wood, they are machine -gunning one ano ther, the y are
dancing, she is rocking a baby to sleep. People are there , the y are doing
something. The sounds I hear mark out the space around me and enable
me to integrate myself into it.
In the dimension of time, music modifies our consciousness of being to
an even greater extent. It is an architecture in time. Ir gives time a density
different from its everyday density. Ir lends it a materialit y it does not ord i-
naril y have and that is of ano ther order. It indicate s that something is hap -
pening in the here and now; chat time is being occupied by an ac tion being
performed, or chat a certain state rules ove r the beings present . The drum-
roll th at resounds throughout the circus ring as the :rapeze artist makes
122 CHAPTER TW•

his death -defying leap is one example. Another is borrowed from my rec
ollecrion of a ritual for the enthronement of a new holder of the power c
Ogun Edeyi (god of iron ) that rook place in a particular village of th
ago -Yoruba territory. Around the sanctuary a crowd of several hundre
people had graduall y collected since nightfall. Noise and bustle . Peopl
coming and going in a darkness broken here and there by the light of kerc
sene lamps. Very graduall y, the lamps go out unnoticed, one after the othe
Total darkness. The night dozes. At about midnight a cry rings out, a ba
of fire hurtles from the sanctuary scattering bright sparks, and sudden]
the drums sound. Their chunder takes poss_ession of the world. The un
vers e has changed. Something is happening . The drums will go on beatin
for eight or ten hours. Dawn w ill drive out the night, day w ill dri ve out tr
dawn, the sun will drive our the coolness of morning, but never for an ir
stant will the drums cease to bear. More correctly, the drum: the litt
drum mentioned a moment ago that is also beaten for oro. It is placed o
the ground and pla yed by rwo drummers at once, facing each other; wh i
rwo more are aligned on the other diagonal, ready to take over. Unlike t~
other drums in its drum section, this little drum does not "speak." c
does it pla y the metronomic role so often ass igned to small drums. It
pla yed more or less quickl y and more or less loudly, sometimes frenzied]
sometimes almost in a doze; in fact it possesses no rh ythmic functio
properl y speaking. E ver yrhing suggests that its role is to maintain a ce
rain vibration in the air, and thus ensure rhe continuity of the action. I
short, to est ablish a different order of dur ation. Or , if one prefers, to brir
about a sort of cr yst allization of rime . In this case it is a matter of crysra
lizing rime for a relati vely long period (half a d ay ), over a relati vely laq
area (the sanctuary an d the space around it ), and for the benefit of a rel:
rively large number o f people (a good twent y protagonists and sever
hundred spectators ). Gi ven such co nditions , the cr ystallization of tirr
cannot be maintained constantl y co the same degree. It inevitably go ,
through periods of ri se and fall. Bur chis is nor of great importance. Wh .
is import ant is th at it should never completel y cease.
Music does nor or ga nize rime onl y at this elementar y level but also at
higher one, thus giving birth co a re al architecture of rime. Possession mi
sic does nor ope rate solel y by means of repetition and accumulation , as
coo often thought. The musical m ottoes are melodic or rhythmic star,
ments, and consequen tly temporal forms. The y are capable of being va
ied and ornamented. In the course of a ceremon y the y follow one anoth<
and thus form sequences char should be seen as the multiple ways of r,
newing and de veloping musical rime, w hich preserves its unit y all d
w hile since the piece s following one another belong to the same genre. E
thus tr ansforming our awar enes s of rime and space, in different wa ys, mi
sic modifie s our "b ein g-in-the -wo rld. "
Music and Possession 123
The state of affective resonance that music - or at any r ate certain kinds
of music - creates in any individual is another aspect of the upheaval it
creates in the structure of consciousness. othing is more laden with emo-
tional associations than music, nothing is more capable of recreating situ-
ations that engage one's entire sensibility. It induces the individual into a
state in which both his inwa rd feelings and his relations with the oute r
world are dominated by affectivity.
Finally, as art, when it is a success, music creates the feeling of total ad-
hesion of the self to what is happenin g. In this sense it again brings about a
tr ansformation in the structure of consciousness, by effectuating a par-
ticular and exceptional typ e of relation of the self to the world.
These very brief remarks were intended solely as a reminder that music
modifies - profoundly, and in several dimensions - the consciousness we
have of ourselves in rel ation not only to ourselves bur also to the world.
These modifications bear essentially upon the perceptible dimensions of
those relations . We shall now see that relations with the gods also fall
within this selfsame category of data.
Ir is clear char possession cannot be understood unless it is set in the sys -
tem of representations of the society concerned . The se representations, in
their turn, must then be brought into relation with the way in wh ich the y
are experienced in dail y life, and in particular with the way the presence of
the gods is experienced in da y-to -day life, outside any form of ceremony.
The least one can say is th at we are ver y ill-in formed about this last kind of
experience. What we do know, hO\\·ever, is that in societies ha ving pre -
sen·ed an archaic way of life, which are precisely those in which posses-
sion cults are frequently found, the individual lives in constant sensorial
contact with nature. He lives in perpetual intimac y with the elements,
plants, animals. For him, the frontier between the animate and inanimate
worlds is extremely vague . Men, beasts, plants, and things - all have souls
or are the receptacles of souls. E very phenomenon is interpreted as result-
ing from the action or presence of a soul. T he visible is constantly animated
by the invisible. For him, the presence of the gods is materialized in all
sorts of objects, localized in all sorts of familiar places, associated with all
sorts of day -to -day activities and concr ete phen omena. Given all this, we
are justified in thinking that he sees the gods as beings who are very close
ro him physically . Their existence does of course become the object of ab-
stract representations on his part, bur everything suggests that he strongly
responds to it on the level of sensory experienc e. I should add that, in
many societ ies , the majorit y of the souls peopling the invisible world are
those of the dead, of dead persons wit h whom the individual sometimes
lives in very close physical proximity (in Fon and Gun terrirory, which is to
say in vodun country, the dead are bu ried in the family dwelling ) and to
whom he refers at ever y occasion.
124 CHAPTER TWO

We kno w, moreo ver , that he often attributes illne ss either to the pres -
ence of a spirit w ithin him who w ishes him ill or wa nts something from
him , or to the theft of his soul, or of one of his souls, by some such spirit. ·
In such a context it seems to me th at nothing could be more sensible, or
more logical, than to interpret sickne ss as the dispossession of one's self by
some othe r who robs you,..if yo ur forces, divides yo u , alien ates yo ur ph ys-
ical or mental capac itie s. Ph ysical illness is an invas ion of the conscious -
ness by sen sa ti ons coming from a pan of the body - a more or less impor -
tant , more or less narrowl y local ized p an depending on the particular
compla int - that is at odds wi th the rest. For an individual who does not
p icture his bod y in terms of the bi olog ical categories of our modern wo rld ,
and w ho does nor int erp ret its dysfunctions as caused by bacteri a or
chemical imbalances, nothing is more natural than to believe that illness
results from a foreig n presence: sinc e one does n or recognize oneself wh en
on e is ill , the ailing limb becomes a foreign body, th e afflicted organs take
on an in dependent existence as th o ugh they had a life o f their own. On the
mental or ne r vo us leve l , sickness - or simply fatigue, irritation, anxiet y,
frustration - also rake th e form of inn er sensations one is receiving from
one's bod y. To be "ove rcome by ange r " simil arly expresses th at the state in
que stion is experienced as a ph ysical modi ficati on of the self. To "feel
d rain ed" very clearl y conveys rhe ide a of feeling dispossessed of one's viral
energ ies . "Etre devore par lafievre" (literall y, to be "devo ured by fever" ) or
"range par la maladie" ("ea ten away by an illness" ) both express that ill-
ne ss is experienced as thou gh there were some ot her living being residing
in one's body. And such expressions exist in all lang uages, so that we can
be certain the y are represent ative of a uni versal exper ience . In the societies
that concern us here, these internal stares are percei ved all the more in-
tensel y at the senso ri al level for the simple reason that da ily life is domi -
nated by manua l labor, which keeps one constantly awa re of one's bod y.
Given these co ndit ions, it ma kes good sense to draw a connection be -
tween such internal stares and the physical presence of one of the souls,
spi rits , or genii, in short, one of rhe di vinities rhar people the external
world and regulate all of its phenomena - good sense that indeed firs in
perfectly well wi th the commonly held rheorv on the gene ral functioning
of the world .
We are thus led to think rhar, in the societ ies conce rning us here , the
ways of being affected by music, of expe riencing rhe presence of th e gods,
and of undergoing illness all de rive, in the sphere of the structure of co n-
sciousness, from one and the same type of relation to the world and to
oneself. Thi s doubtlessly explains why gods, sickness , and music can be
so closely inte rwoven in possess ion cults .
3 Music, Shamanism,
Mediumship, Exorcism

The question of who makes the music in possession rituals and what part
the possessee plays in such music (if he takes a part at all ) led me earlier to
refer to shamanism . The active or passive character of the possessee's rela -
tion to possess ion music is nor in fact fully significant unless one co mp are s
it with the shaman's r elation to mu sic. Thi s is w h y we are now going to ex-
am ine this other aspect of the rel ations between music and tr ance. 1
An in-depth study of sha m an ic music still needs to be written. Since
possess ion is ou r pr incip al int erest here, I shall do no more than skim the
surface of this subject, pausing only when it is imp ortant to clarify jus t
how different things fe in the two cases . After a digression on medium-
ship, which often is simply one aspect of either possession or shaman ism, I
shall move on ro a quite different question, also common ro both; that of
music and trance viewed from the st andpo int of exorcism.

MUSIC AND SH.-\M.-\NIS.\!

By now, I hope it is sufficiently clear th at in possession the relation of the

possessee to music is most often passive.' On rhe contrary, the shama n's
relation to music is ac ti ve, since he sings and drums while he exerc ises his
functions. 3 We must, ne vertheless, ask ourselves if this inve rsion is veri-
fied when in the case of sha m an ism we take into account as we did for pos -
session rhe stage the individual invoh·ed has reached in his career.
In possession we saw that the rel ation of the possessee ro music is totally
pass ive during initiation, and rhar it becomes progressively more ac tive as
he passes from the status of neoph yte ro offic iant. In shaman ism, on the
contrary, the shaman's activ ity as musicant begins from the moment of ini-
tiation . Describing the consecration ceremony of a you ng female shaman
among the Araucan of southern Chile, Me tr aux (1967 , 19 1ff. ) tells u s
how the new initiate, in a state of trance, climbs the rewe- the tree made
into a ladder - which is the "sym bol p ar excellence of he r cond iti on." She
climbs it "dancing and drumm ing."' Having reached the rap, she utters a
long prayer during which she begs for clairvoyance, effec t ive chants , and
the art of drumming. Her behavior is thus the opposite of that of a can -
domble neophyte, for example, who cannot rake the lead in beg innin g a
126 CHAPTER THREE

song until he or she has been an adept for seven years. The Araucan sha -
mans perform as both singers and instrumentalists from the very outset.
Among Asian shamans, on the other hand, full master y of the drum (sha -
manism's musical instrument par excellence ) is not achiev ed until after a
certain length of time has elapsed. In Siberia, among the Tungu s (or
Evenki ), after having made a stick for his future drum, the apprentice sha -
man "remained seated outside for several d ays chanting shamanic chants.
One or n -vo years later, he saw in a dream the reindeer whose skin would
cover his drum" (Vasilevic 1969, 251 ), 5 and it was only after the making
and "animation" of the drum 6 that "the aplfentice shaman was granted
the right tO shamanize ." As a beginner, the Bur ya t shaman has no right to
use a drum; he can only pla y a Jew's harp. 7 Among the Ammasalik Eski -
mos, while the shaman is still a postulant he does not use a drum: "If he
wants t0 provide the rhythm for the chants he learns far away from men,"
he uses "two wooden sticks struck one against the other" (Gessain 1973,
15 3 ). Although the Tungus or Eskimo shaman becomes an instrumental -
ist only gradually, since to play the drum (the geometr ic locus of all the
symbols of shamanism ) is to have mastered all the powers, he is nonethe-
less a singer from the very start, just like his Amerindian homologue, a
fact that distinguishes him radically from the possessee.
The most important differ ence, however, is th at in every case the sh a-
man is the musicant of his own entr y into trance. In other wo rds he goes
into trance not by listening to others who sing or drum for him , but, on
the contrar y, by singing an d drumming himself. All the accounts agree on
this point. And herein lies the great difference between shamanism and
possession. Th e possessed person is never the musicanr of his own trance;
the shaman alwa ys is. For the shaman, shamanizing and musicating are
two aspects of one and the same acti vity. So much so that among the Yar-
uro, in Venezuela, he is called toe wame , wh ich means "musician" or,
more precisel y, "h e who sings and dances ." 8 As for the word shaman itself,
w hich as we saw originates w ith the Tungus , it has now been established
that it derives from the root sam -, "common to all the Altaic languages, "
w hic h "co ntains the idea of dance and leap o n the one hand, trouble an d
ag itati on o n the other" (Lor-Falck 1977, 9 ). In Tungu s there also exist
"three other series of terms expressing the idea of shamanizing" (ibid.,
11 ). 9 T he root of the first isj ajar-, w hich is the "only one that furnishes a
term , jajan, designating the shaman, a term of ancient and frequent us-
age" which o riginally meant "to sing , to pray to the fire for good fortune";
in Chukchi it gav e \.Vayto )ajar , "drum. " The second has two meanings:
"na rrate, sing, hand down the tr adition " and "shamanize" (ibid., 12 ). 10
Among the Tungus, then, language itself identifies the shaman as he who
dances and sings; in short, to express the m atter in general terms, " as a
musicant. But we should note that he is seen as a singe r only with refer-
Music, Shamanism, Mediumship, Exorcism 127
12
ence t0 words, to what is said, and not with reference to the tune or
rh ythm tO which those words are sung. I shall return to this point shortly.
We briefly saw in a previous chapter that this is not true for possession,
since the term used tO identify the possessee has no connection with dance ,
words (whether they are sung or not ), or musical instruments.
As we did with respect ro the possessed person, we must now ask what
role music pla ys (only this time it is the music of the shaman himself) in
triggering the shaman's trance. And let me say right away that this is a vast
problem indeed, and one I shall only touch on.
Let us first talk about the shaman's musical instruments. aturally
one's thoughts immediately turn to the drum, 13 an instrument closely asso-
ciated wit h the im age everyone has of the Siberian or Eskimo shaman .
Eliade believes that if it contributes to inducing trance, it does so for magi-
cal reasons. "It is musical magic 1" that has determined the shama nic func-
tion of the drum," he writes (1968, 149 ). According to him, this magic
can be explained by one of two hypotheses, dep t>nding upon w heth er it
derives from the charm of the drum's sounds, considered tO be the "voice
of the spirits ," 1 ' or whether "an ecstatic experience had been achi eved as a
result of the extreme concentration induced by prolonged drumming."
Wh y "prolonged drumming" should create "extreme concentration" is not
explained . Like~o many others, Eliade is, I fear, falling into the trap of
stereotypical ideas about the power inherent in the physical sound of the
drum. 16 In the case of possession, we were led to conclude that alt hough
the drum is frequentl y used tO induce trance , an y instrument at all can per-
form the same function. And the facts plainly demonstrate that the same
holds true of shamanism, even though the actual variety of choice is less
great.
Among the Tungus, "many groups of Transbaikali have replaced the
drum w ith a pair of canes that fulfill the same functions" (Dela by 19 7 6,
112-13 ). In fact, these canes ser ve as "supports for spirits, " musical in-
struments (the y are furnished with small bells and various kinds of jan -
gles - all highl y symbolic ), and a "means of locomotion" (like the drum,
but by virtue of a partl y different symbolism, the y enable the shaman tO
undertake his journe ys into the upper and lower worlds ). We have seen
th at at the outset of his career at least, the Buryat shaman does not use a
drum but a Jew 's harp. Among the Mirghiz it is replaced by a lute, the ko -
buz (Eliade 1968, 150 ). We know that the Arau can shamans in Chile use a
frame drum th at displa ys curious similarities ro the one belonging tO their
Siberian homologues. 1 ' But among the Amerindians the shaman often
uses a rattle as his musical instrument, at least when he has one, for he can
do without one just as well, as is the case among the Selk'nam (or Ona ) of
Tierr a de] Fuego, who shaman ize without any instrumental accompani -
ment. 18 To end, let us also mention the rather paradoxical case of the Es-
128 CHAPTER THREE

kimos who, unless I am mistaken, all know the drum but do not all use it
in order to shamanize . Alongside an active form of shamanism on the Arc - .
tic coast as far as Greenland, in which the shaman m akes great use of the
drum to induce his trance, there is also among certain "Ce ntral Eskimo" a
more pass ive and contempl ative way of shamanizing (Holtved 1967, 26 ).
The shaman has hi s visions w hen he is seared or lying down. And rhe
drum is not used in these circumstances . Tri ggering tr ance in shamanism
thu s appears to be no more linked to music al instrument s in gene ral, and
to the drum in particular , th an it is in pos session.
On the other hand , in order to tr igger trance, shaman ic music , in some
cases at leas t, does m a ke use of rhe simultaneous crescendo and accele -
rando that , as we ha ve alre ady seen, are at work in possession. Shiroko -
goroff, w ho in his monumental stud y on T ungu sian shamanism stresses
the "except ion al psychological po we r of the music al pan of shamanic per-
formances" (1935, 327 ), refers on several occasions, an d in different
cha pter s, to the in cre ased tempo and vo lume of both drumming and
chanting, which he says strongly contribute to inducing trance ' 9 ow ing to
the resulting intensific ation of emotions. Th is appli es both to the shaman's
incorporation of his auxiliar y spirit s' 0 and to his journey into the lo w er
world." According to him , the shaman's aim "is to put him self int o a sta re
of ecstasy, which is achie ved by va rious means - rh ythmic drumming wi th
an incre ase of rem po, singing, 'da ncing' and even drinking wine" (ibid. ,
339 )." Simil arly, the au dience 's entr y into trance - for it can happen that
the spectators go int o tr ance as well - is achie ved by the repetition of re-
frains, "w ith graduall y incre asing emoti on chiefly produced by an increase
in tem po, int ensity, and expression" (ibid. ). But let us first stress that ac-
celerando and crescendo do not always have the effect of triggering tr an ce.
Shirokog oroff himself tell s us that "the increase in the rem po of drumming
and the singing and dancing become a ritu al wh en the shama n does it for
influencing the patient , bur does not himself feel to be in ecstasy (ibid. ).
Second, so metimes different, if not diametricall y opposed, m usical meth-
ods are used . Describin g the sha m an's prep aration s for achieving his stare
o f "doubling, " Shirok ogoroff says th at "a t the beginning the shaman
drums , continuall y incre as ing and decre asing the temp o and the intensit y,
with a definite rh ythm empirically disco vered , in order to produce a ph ys-
iological and p sychic state in w hich a sugge stion - the coming of a spirit -
made by the sha m an himself , ma y h ave an immediate effect of 'dou -
bling.' " (ibid . , 363 ). Sometimes, then, tr ance is trig gered by the constant
acc eler ati on of tempo, at othe r time s, on the con tr ary, by an alte rn ation of
acceler at ion an d deceleration . It seems, too, that what ever its rh ythmic
fo rm m ay be, the sound of the drum is thou ght to possess within itself the
powe r to call upon spi rit s. 23 And finall y, the "co ntent of the text of the sh a-
m an's songs" also pl ays a part (ibid. , 330 ).
Music, Shamanism, Mediumship, Exorcism 129
As wi th possession, then, in shamanism the music al means used to trig-
ger tr an ce are va ried. Moreo ver , the y operate onl y w ithin the framework
of an ove r all situation. For in fact, al though the Tungus shaman may cer-
tainly be rega rded , acco rding to Shirokogoroff's acco unt, as the principal
mu sica nt of the shamanic seance, he is no netheless supported in that role
bo th b y hi s ass ista nt , who frequent ly cakes ove r the drumming, and by all
chose w ho are no t o nl y present but also participate in the seance. The y all
help him by ch anting and maintaining th e rh ythm , wh ich has the effect of
intensifying his trance. Several sham ans rold Shiro kogoroff (ibid., 363 )
tha t the y w ere un able co perform without an audience . "All people pre -
sent," o ne of chem rold him, "helped me go co the lower wo rld ." In o cher
words, the shama n , just as much as the possessed pe rso n , bur in a differ-
ent way, mu se have an au dience.
Among the Selk'nam of Ti er ra del Fuego, the shaman (male o r female )
w ent inro trance by singing (aga in witho ut instrum enta l acco mp a niment )
and giving in co frenzied ph ys ical activ ity, dancing , leaping into the ai r ,
striking che earth violentl y with his feet and fists,2' repea tin g, all the
while, the same in canta tions aga in and agai n. \Xlas chis repetiti on also ac-
co mp anied by an accele ration of the rhy thm and an incre asing intensit y of
che singing? We just don't know. We ca n be quite sure, howe ver, th at the
techniques used ro induce tr an ce va ry- or va ried - a great deal from one
country (or continent ) co another. Among the Selk'n am, w ho did not use
drugs ac all, trance was achie ved by ent irel y n at ur al me ans of auroex cica-
cion , but recourse ro artificial me ans is co mm on in bo th Asia and Amer -
ica . Among che Yarur o of Vene zuel a, S. Dre yfus cells us, "trance is, if not
induc ed, ac leas e fac ilit ated by the abso rption of a ver y large quantit y of
stimulants {fermented beverages, alco hol, tobacco )" (1954, 3 )." In Sibe -
ria, among the methods used by the Yaku t sha man ro "reach the altered
st ate ," E. Loe-Falck (1968, 2 : 264 ) cites m ycoph agy as "the surest meth -
od, provok ing hallucinations aucomaci cal ly."
But lee us get back co music. T he ex teriorizacion o f rhe sha m an's so ul
during his journey ·'manifests itself in two diametric ally opposed ways,"
E. Loe-Fa lck cells us , "t hrough catalep ti c tr ance and throu gh dr am atic
tra nce" ( 197 3, 9 ). T he music is clearly no t ac all che same in che cwo cas es.
During catalep tic tr ance, the shaman's body is "rigid , no mo re than an
empty envelope, abando ned by che so ul, ,,vhich has left co acco mpli sh
some mission," and it is rhe shaman's ass istant who is che mu sicanc. Dur-
ing drama tic tr ance, on the contrary, the shaman describes wha t he is see-
ing during his jou rney into the upper o r lowe r wo rld and recount s his ad -
ventur es by chanting and beating his drum, thu s providing a tru ly
thea tric al performance, or, mo re acc ur atel y perhap s, a one -man show,
which includes epi sodes drawing upon the most varied mu sical styles :
songs, recit ati ves, invoca ti o ns , spoken p assages, dialogues, imita ti o ns of
130 CHAPTER THREE

anim al cr ies or sou nd s of nature and other ono m ato poei as, and vo ice dis-
guises. 26 Dr amatic or epic moments altern ate w ith comic passages . The
drum pro vides a constant support for th e sung action, underscoring the
vocal effects and pro viding links between one episode and the next. Its
function is someti me s rh ythmic, and to be sure, it always accords w ith the
character of the particular song or narrative - hurried, panting, solemn ,
peaceful, and so on. At other times, its function is one of pure punctu a-
ti on, and the meaning of th e drummed phrases is some tim es symb ol ic ,
sometimes descripti ve, somet ime s simpl y a kind of signal. At least, this is
what one gathe rs from one of the rare ava ilable recordings of shama nic
music origin ating in Siberia. 27
At the anti podes , among the Wayana Indi ans of Guya n a, 28 another re-
cording suggests that, all things being equal, we are also dealing w ith a
theatralization of the shamanic seance that is rather compa rable to the Si-
berian example. Since there is no musical acco mp ani ment proper, the
sung actio n is supported by "sk illful sound effects" (Hur ault 1968, 4 ) pro -
duced by very rudimentary means. Among the Yaruro of Venezuela, the
shama nic seance held for the purpose of curing a sick person also begins
w ith a narration of the journey made by the shaman 's soul "into the rea lm
of Kum a, the mothe r goddess " (Dre yfus 19 54, 1 ). The shaman sings, ac-
companying himself with a gourd -r attle "engra ved wi th figures and signs
representing inhabitants of the supernatural wo rld " an d depicting, "ar -
ranged around Pwa na, the great wate r serpent, husband of the anthropo -
m orphic goddess Kuma, the tribe 's male and female ancesto rs dancing in
the dvvelling place of the dead. " As he sings, the audience replies in chorus .
Th e "auxiliary sp irit s" are incorporated in the course of thi s jou rne y.
When the shaman leans over the patient and "tries to extirpate the sick -
ness ... by aspiration and suction of the painful parts," he stops singing,
and his assistant then takes over this function for him. Ha ving successfully
"extracted the pathogenic element , he utter s a cry, throws himself back-
ward, falls to the ground, a nd remains unconscious for some moments"
(ibid., 3 ). Thi s cris is, which is directly related to a particularl y dram atic
moment of the heal ing proces s, does not seem to be linked to any simulta -
neous dramatization of the music.
Although the pantings, cries, whistles, sighs, hoarse breathing, yelp s
("wh ich are the sign of the spi rits' presence" ), and the occas ion ally qui ver-
ing vo ice of the shaman do cre ate a certain theatr alizat ion of the seance
and introduce some va riet y into the dev elopment of the sung actio n , fo r
the listener who cannot understand the language and is consequentl y un-
ab le to follow wha t is being sung, this act ion would tend to seem very
monoronous, despite the introduction of new themes that fo rm a ki nd
of suite. If I had to characterize the style of this seque nce of songs in a
single wo rd, I would sa y the y are incanta tions (though this is an entirely
subjective eva lu ation , and based mainl y on the monoron y I mentioned, on
Music, Shamanism, Mediumship, Exorcism 131
the one hand, and on a certain qualit y in the intonation of the singing on
the other ). This is precisel y the term Claude Levi-Str auss uses (1958, 205 )
in the case of the famous shamanic "so ng" whose purpose is to "a id a diffi-
cult birth " and that comes to us from the Cuna Indians of Panama. This
long narrati ve, whic h undoubtedl y relates the most extraordinary sha-
manic jo urne y imaginable, inspired Le'vi-Str auss to draw comparisons
between shama nism and psychoanalysis th at are wo well know n to be
repe ated her e. In his interpretation of the "symbolic efficac y" of this "gre at
magico-religious text," he make s no mention of the connections that
might exist between the shaman's song and his tr ance. And he does not do
so for two re asons; first, because this was not his subject, and second ,
because the shama n is not, or at any rate does not seem to ~e, in a trance.
W hereas "transition to an altered state" is not nece ssa ry for the shaman to
undertake hi s journey, which is required fo r hi s curative action , we are
told that clair voya nce , w hich the Cuna consider as the shama n 's "inn ate
tale nt ," is abso lutel y indispensible (ibid. , 206-7 ). Among the Buryat ,
R. H am ayo n tells us (1978a, 63 ), w hen the shaman goes into the st ate we
have termed "tr ance," they say he "ent ers into a visionar y srare." When he
desc rib es his journe y he is, aft er all, describing what he is seeing. Clair -
voyanc e o r journey, what he ch ants has a meaning, and it is the meaning
of thi s sung text that is important . The re aso ns why the words ha ve been
set to mu sic probabl y ha ve much more to do w ith the difficult y of
memorizing a very long text , or with the need to render ir expressive or en-
dow it w ith a certain m agical effect, th an w ith any attempt to engender
tr ance .
One of the great differences between po session and sha man ic music
lies in rhe fact rhar the latter is often endowed w ith magical power, w here-
as the former is not. It speaks, it identifies, it induces move ment. !rs efficacy
is of a mechanical order. Its effects are of a p ractic al o rder: it summons the
gods, makes the dancers whirl, cre ates contact between men a nd gods,
arouses emotion. It has no other powers. The shaman's music, on the con -
trary, claims, in certa in cases at least, to transform the world. To use Levi-
Srrauss's \\'Ords (196 2, 292 ), ir seeks t0 intervene in na t ural deter minism
in o rder tO modify irs course . Ir is in this re spe ct rhar iris m agical : through
the power of incanta tion whe n it is meal an d when rhe words and their
musical shaping are w hat count, or th rough the power o f so und w hen it is
instrumental, the sym bolic nature of the instrum ent being then as much at
work as the spell of rhe sound. Of the thirt y-four shamanic songs Anne
Chapman (discograph y, 1972 ) collected among rhe Selk'nam of Tierr a de!
Fuego , sixteen 2 9 relate to notions of journey and magical power th at con-
cern not only heal ing bur also such act ions as conjuri ng a lunar eclipse ,
aiding whale hunters, or pre venting rain through rituals. 30
Inc antator y : such is the main char acteristic of shama nic mu sic. But
\\·hat is an incantation? Thi s is a vast problem I can onl y wuch on here.
132 CHAPTER THREE

In his description of the six basic functions of verbal communication,

R. Jakobson defines "the magic, incant arory function" as "chiefly some
kind of conversion of an absent or inanimate 'third person' into an ad-
dressee of a conative message" (1960 ). The third of the three examples he
uses to illustrate this definition is taken from the Old Testament (Joshua
10:12 ): "Sun, stand thou still upon Gibeon; and thou, Moon, in the
valley of Ajalon. And the sun stood still, and the moon stayed." This type
of text, which juxtaposes conative messages emplo ying, first, the conver-
sion of a "third person," and, second, referent ial messages also concerning
the second or third person, but in which "the stress is at the same time laid
on the message itself ... which is characteristic of the poetical function
of language" [(ibid. )], corresponds fairly well with what we know of sha -
manic incantations. As for the musical character of the words, that is, the
singing itself , it seems that its particularities are, first, ro be very repet-
itive; second, rouse a type of vocal delivery nearer ro the spoken than the
sung; and third, to make use of unusual vocal effects-cries, onomato-
poei as, imitations of all kinds of noises - all of which combine to give sha-
manic singing a distinct position with in the musical system characteristic
of the population concerned.

In a previous chapter/ ' I tried to isolate the difference between shaman -

ism and possession and came to the conclusion that ir could be expressed
by a series of three oppositions: journey to the spirits / visit by the spirits;
control over the spirits / submission to the spirits; volu ntary trance /
involuntary trance. This triple oppo sition could be further condensed into
only o ne: acting / undergoing. Shamanism appeared to be, if one ma y say
so, essentially acted, possession as undergone.
To use Pouillon's terms ( 1972, 94 ), "the orientation of the relation " be -
tween subject and trance appeared to be diametrically opposed in the two
cases. We just foun d a similar opposition in the relations between subject
and music . In possession, as we saw, music is made for the possessee . In
shamanism it is made by the shaman. The latter is a musicant, the former
is musicated. This opposition musicant / musicated should be considered,
therefore, together with the three previous ones, as a factor contributing
to the distinction to be made between shamanism and possession.
Shamanism or possession: in both cases the imaginar y content of trance
is of course totally socialized. By this I mean that the representations un-
derlying this content are shared by the group as a whole . The musicant /
musicated opposition means that the shaman is a musicant since he im-
poses his imagin ary world upon the group, whereas the possessed person
is musicated since the group imposes the imaginary content upon him.
Moreover, in possession the subject goes into trance because he changes
identity; in shamanism he goes into trance because he changes worlds.
Music, Shamanism, Medi umsh ip , Exo rcism 133
The ro le of m us ic in sha m anism is thus d ou bly differ ent from th e one it
plays in possession: firs t , because music is ex perienced di fferently depend-
ing on whe t he r one makes it onese lf or w het he r others make it ; second,
because it serves two diffe rent imaginary cons tructs (o r, if one pr efers ,
rwo diffe rent ideo logies ) of tr ance. In possess ion, utter ing m usica l m ot-
toes associated with the identificatory aspect of trance plays an impo rt ant
role in triggering chat trance. In shamanism, it seems chis function is pe r-
formed by the incantatory character of the music. But in shamanism as in
possession the facts are quite variable. T he ava ilable accounts lay em p ha -
sis somet imes on the shaman's physical agitat ion, 32 so m etimes on the
chanted summons, and sometimes on his use of the drum. In mainta ining
the trance scare, as opposed to triggering it, music constitutes in every case
an essential factor guaranteeing continu ity during the shamanic journey.
Among the T ungus, after the shaman has entered his trance by ch ant ing
and drumming, his assistant cakes the "d rum and contin ues the d rum-
ming, both for maintaining the shaman's stare of exras y and for contro l-
ling the behaviour of the audience," Shirokogoroff cells us ( 1935, 329 ). In
face, the journey is pursued from episode to episode through a musical d i-
mension, and it is both the words and their musical sha p ing chat endow
the singing, when necessary, with its incantato ry and magical va lue .

MUSI C AND MEDI UMSH IP

Shamanic and possession trance have in common the fact chat they bo th
frequently lead to divinator y practices, as we saw earlier. 33 Insofar as sha-
mans and possessed persons behave like med iums who prophesy or are
consulted, I ha ve nothing to add here about the relation between the ir
trance and music. Bue it can happen chat the exercise of mediumship, or, if
o ne prefers, of divinator y trance, acqui res a certain autonom y, so that it
muse then be distinguished from shamanism or possession in the stric t
sense of the term. From the standpoint of its relat ions with music, t he situ-
ation is thus a somewhat hybr id one. I shall now examine chis aspect of
the matter.
With a view ro distinguishing between possession in the usual sense of
the term (spirit possession ) and medium istic possession (spirit medium -
ship ), R. Firth (1969, xi ) points our chat, in the first case, the possesse d
person's behavior does nor necess aril y "transmit a particular message to
othe rs," whe reas in the second "co mmunication is emphasized ." The spir -
it respons ible for mediumistic possessi on has something to say to a n audi -
ence. And obviously when the divinity is speaking through the medium 's
mouth, he should be heard clearl y. This means that there ca n not be any
music at the same time. I will only cite three examples. In Porco - ovo
(Benin ), when rhe priest of Hwonse, a vodun who utters the oracles,
134 CHAPTER THREE

prophesies in public or answers questions put to him (something I have

witnessed on several occasions ), one could ha ve heard a pin drop . In his
description of a mediumistic trance seance among the Shona in Zim-
babwe, Paul Berliner (1975 /7 6, 137 ) tells us that as soon as the climax
marking the possession of the medium by the sp irit is over, the mus ic stops
and the participants all sit down to listen to what he is going to say. Later,
when the spirit has left, the music resumes . And we have already seen in
an earlier chapter that when the Tibetan oracle begins to prophes y the
music stops .
The medium 's entr y into trance, on the other hand, often requires m u-
sic . When it does, two possibilities exist. Either the music is pro vided not
by the medium but by musicians or musicants grouped around him, which
seems to be the case most frequently, or else the medium himself is the mu -
sicant of his own entry into tr ance. In other words, in the first case events
follow the pattern of possession and in the second that of shamanism. Let
us lo ok at a few examples of both kinds, beginning not with identificator y
possession trance - the trance state centr al co this book-but rather with
the type of trance I proposed to term inspirational or communial. Whe n
the Hebre w prophets were in thei r state of prophetic frenzy, the y were ac-
companied by a va ri able number of musicians. 30 Elisha we nt into trance
and prophesied wh ile he was listening to music. ;; The se prophets are never
described as their own musicants. During the religious wars in the Ce -
vennes (France ), it was quite usual for people to go into trance and proph -
esy during the ass emblies in which there was singing; but often the y were
also "seized by lnspiration" 36 and prophesied without any music w hatso-
ever . Eve rything depended on the circumstances, but I shall not delve any
more deepl y into this aspect of divinatory trance. Prophetism is a vast sub-
ject that extends far beyond my present aims. Let us thus move on directly
to possession. ' '
In Tibet the music is provided exclusivel y by the medium's assis tants
and consists, as we ha ve seen, in sung invocations accompanied by drums,
bells, or cymbals, and sometimes by a horn . At least this is wha t happened
in the case of the ''St ate' ' or acle; I shall cite a different case later. Black Af-
rica, \\·here possession and medium ship frequently coexist wi thin the
framework of the same religious system, provides a perfec t terrain for
co mparing the relations of these two aspects of religion with music under
particularly favorable conditions. And the comparison is very instructive,
as \\·e shall see. In Zambia, amo ng the Tonga (Colson 1969, 76 ), before
entering his hut to answer questions put by those consulting him, the ba-
sangu medium m ay sometimes go into trance publicl y while others clap
their hands and sing to summon the sp irit s. As a general rule, however, he
does not enter into trance until he is inside the hut, and does so without
any stimulus other th an fumigations. On the contrary, possession in the
Music, Shamanism, Mediumship, Exorcism 135
usual sense of the term - a state due to the masabe (a qu ite diffe rent ca te-
gory of spirits ) and one that has given rise to a cult mainly co ncern ed w ith
th_erapeutic activ ities - is always accompan ied by drums and sing in g. The
difference is thus very marked. Contrary to w h at we find to be th e case in
possession, the medium 's trance is not necessar ily accompan ied by mu sic,
and if it is, the music is of no great importance. Among the Nuba of Ko r-
dofan, during mediumis t ic "seances" organized for "clients " (Nadel 1946,
36 ), the medium 38 goes into trance without mus ic of any kind.
Among the Kalabari of igeri a (Horton 1969 ) different forms of pos-
sessio n are found, some due to the spirits of the group's founding heroes,
others to those of water gen ii. The y all, however, give rise to two categor-
ies of manifestation; possession dances integrated into a va riet y of rituals
on the one hand, and mediumistic seanc es on the other. In the case of the
possession dance (oru seki or dance of the spirit ) assoc iated with the cult
of the founding hero, the priest's tr ance is triggered by a deafening din pro -
duced by drums, praise songs, and the sound of machete blades struck
violentl y together to a very insistent rhythm (ibid., 19 ). In the case of the
mediumistic seance oru bibl n'ekwen (speaking through the spirit's
mouth; ibid., 25 ), the medium, who is a different person from the priest
who dances the oru seki, goes into trance afte r the clients (who have come
to consult him ) ha ve summoned the spirit by means of a n invoca tion and a
libation of gin. Perhaps this invocation is sung; the account is not clear on
this point, but there is nevertheless no mention of music. There is thus a
total contrast between these two entries into trance. Moreover, the sa me
contrast can be observed in the posse ssion attr ibuted to water genii: when
it rakes the form of possession dance there is music (ibid . , 30- 33 ), and
there is consequently every reason to believe - although the rexr does nor
sa y so - that entry into trance has musical accompaniment; when it takes
the form of consultation with a medium there is none , but I shall return to
chis point a little later.
Now let us go on to the second case, in which the medium triggers his
trance himself by acting as his own musicant. The most str iking example is
provided by the Sukuma of T anzania (T anner 1955 ). Among them, the
medium is possessed by the spirit of an ance sto r. He goes into trance in or-
der to repl y to questions put by people w ho come to consult him, diagnose
the causes of their misfortunes or illnesses, and prescribe a remed y for
them. Seated on his stool, he begins turning his rattles in front of him,
close to his head, and then starts to mutter incre as ingl y loudly. This is how
he triggers his own trance. When the moment comes, he suddenly stops
shaking the rattles and begins to talk . "The stopping of the rattling is said
to be done by the ancestor spirit, and not under the direct vol ition of the
magician" (ibid., 276 ). This is because the spirit has been disturbed by the
noise of the rattle and therefore hurries to "assist i :1 the seances in order
136 ' CHAPTER THREE

that the noise ma y srop ." This medium can also, if the need arises, trigger
possession in the patient con sulting him , still, of course , in the hopes of
finding a remed y for his ills. In this case he doe s not enter into trance him -
self. H aving whistl ed tO su mmon the spirits, he begin s tO shake his rattles
close tO the client 's ears, slowly at first but then more and more quickly.
"The slow crescendo of the rattling allows for the grad ual arrival of the
spirit concerned , since if it arrived too quickly it would have a harmful ef-
fect on the patient" (ibid. , 277 ). Thi s, then, is a doubl y interesting case: it
demons tr ates that the very same person is capable of tri ggerin g tr ance
bo th in hi m self and in others; it shows that the musical means are the sa me
in both cases, but that the interpretation of the effects they produce differs
in each.
In Niger ia, among the Kalaba ri mentioned earlier, when a wo m an pos-
sessed by a water genius acts as a mediUI,1, she triggers her own tr ance
o nce her clients are there by placing offerings on rhe alta r of the spirit pos -
sess ing her and "by calling upon it to mount her " (H orton 1969, 29 ). The
reverse is the case, as we have just seen, when the medium is possessed by
the spir it of a founding hero: the tr ance is triggered not by the medium
himself but by the clients who are consulting him. In both cases, however,
the means are the same: summo nin g the spirit and giving offe rings; there is
no music, unless perhaps the summ ons is sung, somethi ng we d o not
know. Let me sum up the situa tion amo ng the Kalabari: possession in the
usual sense of the term is most o ften triggered by music, and by music th at
is never made by the possessed themselves; mediumistic possession, on the
contrar y, is triggered without the aid of music, or at most by chanted invo -
cations, an d entr y inro tr ance can be induced either by the med ium's cli-
ents or by the medium himself.
Benin (formerly Dahome y) supplies a slightly different exam p le. Th e
Gun of Pon -Novo have a vodun, called H wonse, whose principal func-
tion is ro deliver or acles . He is also consulted in order to establish the truth
and settle legal wr angles; occas ionall y he is thus summoned to assist the
king in making judicial decisions . Like all vodun he has his an nual feast, and
on one of these occasions (in December 1964 ) I was able to observe him
exercising his functions. For this feast the ceremonies begin with a solemn
reopening of the sanctuary, which had been closed for seve ral weeks , fol -
lowed by an exhibition of the vodun and an offe rin g of the first yams. All
this rakes place, naturally enough, in an atm osphe re of grea t rejoicing. In
rhe open space bern·een rhe sanctua ry entrance and rhe houses facing it, a
small band of drums, rattles, and bells plays music so that the fai th ful and
the local people can dance . The final da y is reserved for a great procession -
al during which the vodun emerges from his temple and is carried around
the sanctuary p recincts on rhe head of the high p riest. This is followed by a
public seance during which Hwonse speaks through the mouth of his
bearer.
Music, Shamanism, Mediumship, Exorci sm 137
How is the medium's trance achieved? First he w1shes his head w irh lus -
tral water. Then he places the vodun, w hich is rather hea vy, on top of h is
head. A drummer then plays on the agbla, an instrument used exclusively
for such cult purposes. A note in passing: this type of drum, which is not
specific ro H wo nse bur is found in che worship of ocher vodun, is no t
meant ro produce very loud sounds . The drum itself is shallow and entire -
ly open beneath, while the skin is relatively loosely stretched. It is pla yed
wi th the hands, by gently tapping the membrane, just hard enough for one
to hear the calls, greetings, and mottoes the player addresses, in drum lan -
guage, robe sure, to the divinity. There is no sing ing. While the drum is be -
ing pla yed ro summon the vodun, the pr iest also utters mottoes (mimlan )
and speaks to Hwonse wi th the help of words "that he alone knows and
that force Hwonse to come and incorporate him." At least , this is what I
was rold when I made inquiries following the event, since circumstances
prevented me from being in the sanctuary precisel y when the med ium
went inro trance.
For our present purposes there are two facts robe noted here. First, that
wo rd s, spoken or drummed, play an import ant role in this entry into
trance, bu t that music plays almost none, since ic would be improper co
view the agbla as a musical instrument, at lease as it is played in chis situa -
tion. Thi s drum is talking. If need be, one might say ic "mus icaces" a bit,
but one cannot re ally say ic "makes music." Second, rh ac H wonse's medium
does play a role in his own entry into trance, but he is not alone in bringing
it about . Alth oug h he himself pronounces the wo rd s chat fo rce che vodun
co incorporate himself, ic is equally necessary for che drummer ro address
words of praise co the divinity through the agbla . With respect co how the
trance is triggered, then, and co the role played by the medium himself in
chis process, che example of H wonse illustrates extremely well what I sa id
earlier abou t the hybrid situation chat mediumship, in its relation tO mu -
sic, occupies between possession and shamanism .
Lascl y, from Tibet \\·e have an ex ample of a medium triggering his own
trance by ac ting as his O\\·n musicant; buc this time the process is presented
as being far from usual. As we Sa\\. in che case of the Stare or a cle, trance is,
generally speaking, triggered by music played by che medium 's assistants.
Bue 1ebesky-WojkO\\·icz (1 956 , 437 - 38 ) cells us char another diviner,
whom he kne\\. ver y well personall y, confided chat he sometimes proceeded
in a different way : "If he is asked co deher a prophecy for one of his cus -
romer and there is no priest present who would be able co chant the neces -
sa ry invocation, he will sing the prayers himself. H e assured me chat chis
method, roo, proves most effective, and char before concluding the chant,
he looses consciousness and falls inro the tr ance . Should, howe\·er, the
first prayer have no effect, then he will chant once or cwice aga in." Insofar
as it is self-induced (like those encountered by T anner among the Sukuma ),
this trance is an exception among Tibetan mediums. In all oche r respects,
138 CHAP TER THREE

howe ver, it conforms with the general model of Tibetan mediumistic

trance, which exhibits a very different aspect from the one seen in black
Africa.
As we have seen, 39 the medium in Tibet prepares himself by meditation
and purification, by fast and abstinence. R. A. Stein (1962, 155 ) de-
scribes his beha vior while in trance as follows: "Noc only is his face trans-
formed - it swells and becomes red, the eyes become bloodshot, the
tongue chickens and hangs out, ecc.- buc he also shows proof of superhu-
man strength, supporting an extremely heav y helmet, bending swords,
ecc." Thi s trance, we muse remember, owes absolutel y nothing to the
drugs. ebesk y-Wojko witz obse rves char "medical authorit ies who exam -
ined my material expressed the opinion char rhe trances of the oracle priest
may be partly induced by cecany, the attack being initiated by controlled
respiration" (ibid., 440 ). This would impl y chat trance is produced by vo-
luntary behavior using a form of self-control , an ascet ic practice that is
much more reminiscent of shaman ic practices than of those used in pos -
session. Nebesk y-Wojkowitz himself devotes a chapter to what he calls
Tibetan shamanism, and in it he compares Tibetan religious practices
(particularl y chose concerning the oracle ) and Siberian shamanism. Per -
haps we could sa y chat whereas the possession of the Tibetan oracle, seen
as a religious practice, falls with in the domain of possession, the corporeal
technique he uses in trance falls with in the domain of shamanism . The Ti-
betan oracle would thus provide a particularl y striking example of a situa -
tion intermediate between these cwo major forms of trance religion.
Moreover, whereas the Tibetan medium can onl y occasionally trigger
his own trance, the "true bards" of Tibet, who go into trance before sing -
ing, do so as a rule. In o rder co encer into trance and "summon a hero into
himself" (Stein 1952, 156 ) the bard "begins by purif ying himself with a
pra yer and by singing a description of the hero ." Thi s is thus a self-induced
trance, one chat invo lves possession rather th an shamanism, since.as Stein
writes, "the god of the epic comes down [and ] falls" onto the bard. On the
ocher hand, the 3Ct of singing while in tr ance, which is characteristic of
shamanism, chus seems to be characteristic of a certain form of possession
in Tiber. In chis respect, it wou ld be interesting co knO\\. how things stand
throughout chose areas in which shaman ism and possession coexist, par-
ticularl y in As ia among th e proto-Indochine se (Ponier 1973 and Condo-
min as 19 73, 141 ).
All the cases o f medium ship we ha\·e re v;e\\·ed thus far invol ve at lease a
minimum of music, for o ne muse assume, even when the authors do not
say so explicitl y, that the invoca ti ons or summons addressed to the spirit
muse be chanced to some degree. To what extent was chis also true of the
Pythia of Delphi, chat m ost famous of all mediumistic possessions? "How
was the 'enthusiasm ' evo ked by ancient writer s induced?" asks]. Defradas
(196 8, 192 ). On chis point, author s are quite di vided (cf. Dodds 1957 ,
Music, Shamanism, Mediumship, Exorci sm 139
70- 7 4 ). J. P. Vern ant ( 197 4, 11 ) has recently commented on our stat e of
ignorance as to the workings of "the oracle par excellence, the oracle
whose fame and prestige eclipsed all ochers, and whose authority, from
the eighth century onward gradually extended ove r che whole Hellen ic
world : the Delphic oracle." In order to answer the questions put to her,
the Pychia either went into a "true scare of trance," and spoke in so unintel-
ligible a manner chat her answers required "prophets" to interpret chem,
or else she "answered the questioner directl y" in which case we muse sup-
pose she spoke "in cold blood." In either case there is no quest ion of music
being invo lved. And even supposing char music was used in order to trig-
ger the trance, it must certainly have bee n limited to a few chanted invoca -
tion s, at the very most. In any case, we can be sure that unlike the "te les-
tic" trance of the Corybantes,4 ° which always required music, and a great
deal of music , the "mantic" trance of the Pyrhia, on the contrary, required
very little or, more probably still, none at all.
The relations between divinator y trance and music are thus quire va ri-
able. The fact remains, however, that if the medium is a shaman then in
every case in which music is involved he is the music ant of his own entry
into trance, but if he is a possessed person, then he is in general musicaced.
Ir can happen, however, albeit less often, chat the possessed medium too
can be a musicant, and in chis respect his situation then becomes more like
chat of che shaman. Nor because his trance changes in content - the visita-
tion of the spirit never becomes, as in the shaman's case, a journey to the
spirits - but because the relation of the possessee to the spirit possessing
him is no longer characterized by submission, nor even, quire often, by
identification, which is the defining factor of identificator y possession.
The Sukuma medium in Tangan yika, whose trance T anner desc ribes as
self-induced, ultimately is a professionally possessed person, and his con-
stant relations with the spirits who come and inhabit him in order to speak
through his mouth have given him a greater or lesser degree of mastery
over chem. Like the shaman, or almost so, he becomes master of the spir -
its, and thereb y master of his own trance. An y gradation between the two
extremes is of course possible, ho weve r, and chis explains why the rela-
tions of medium ism to music are so var iable.
Thi s va riabilit y also results from mediumship s vast va riet y of forms
and techniques . The latter are doubtless attributable, at lease in pare, to
the fact char mediumship is often practiced in a completel y private way, in
conditions favorable to indi vidual initiative and consequently very likely
to encourage change .

THE CASE OF THE B USHMEl\"

Among rhe Bushmen," ' music pla ys a determining role both in tr ance,
wh ich is itself linked to dance, and in che treatment of illnesses . Tr ance,
140 CHAPTER THREE

dance, and the treatment of the sick amo ng these people all fall into the
category, not of possession, but, as we shall see, of shamanis m. This topic
is of such interest from the standpoint of this book that it merits fairly de-
tailed examination here.
The Bushmen attr ibute illn esses to the will of the creator god, who dis -
tributes good and ill fortune among humans through a secondary god and
his messengers, the spirits of the dead (Marshall 1969 , 350 ). Tre atment
consists in attempting to drive rhe sickness out, together with its agents,
by using the "supernatural por~ncy" (nl um, which Marshall terms "medi-
cine" ) wielded b y their "medicine men," a potency or a power that also re-
sides in their "medicine songs" and in their "medicine plants. "' 2 This power is
usually possessed by men, ' 3 and medicine men make up more than half of
the group's male popul ation; but women can also possess it. In o rd er to be
effective, this power must first be activated. It is awakened by singing
medicine songs, then heated by manipulating glowing coals and by danc -
ing. "E ventu ally rhe n / um becomes so hot that it boils," and "it is at its
strongest when medicine men are in trance and the songs being sung with
utmost vigour ." Ar this point the healing power reaches irs pe ak of efficacy
(ibid., 352 - 53 ). The medicine men exercise this power eithe r in the course
of medicine dances in which rhe whole group rakes pan - and occurring
frequently though not nece ssa rily in connection with any particular ill-
ness'" (ibid., 354 )- o r during healing seances th at are more pri va te in na-
ture and a imed spec ificall v at the rre annent of a particular illness . To sim -
plify, one might say that rhe first involves preventive medical practices
w here as the second in volves curative pr actices .
Broadly speaking, roles in the medicine dance are distributed as fol -
lows : the wo men sir in a circle around a fire, forming a compact group ,
and sing the medicine songs wh ile accompanying themselves with loud
handclapping; rhe men dance around them in single file, gradually go into
trance one after another, then exercise their healing power, both for the
affl icted and for the group as a whole.
In the medicine dance, trearmenr of the illness involves two stages . The
first consists in capturing and driving out the illnes s: the medic ine man, in
a trance, massages the sick person's bod y, anoints ir with his own perspira -
ti on, lays h ands upon him, thus drawing the illness out into his own bod y;
he then rids himself of it by ejecting it violentl y in the direction of rhe sec -
ondary god and his messengers, rhe spirits of the dead (ibid . , 370 ). The
seco nd st age consists in driving away the spirits themselves by hurling
sticks and insults ar rhem , and sometimes, too, though more rarely, by tr y-
ing to conciliate them by calling for their help. When he is entranced, the
medicine man is in fact seeing the secondary god and the sp irit s of th e
dead, and is therefore able to address them directly. When he is at the
deepest st age of his tr ance, "in half death," his soul (or rather one of his
rwo souls, it seems ) leaves him to go meet the secondary god (ibid., 378 ).
Music, Shamanism, Mediumship, Exorcism 141
In olden times, the Bushmen say, there were medicine men who act u ally
met the great god himself, when he let down a rope from heave n so the y
could climb up to him. Clearly then, we are dealing on all cou nt s with sh a-
manism, not with possession.
R. B. Lee (1968, 37 ), in his article on the sociology of trance among the
Bushmen , emphasizes the different roles played by the two sexes in the
medicine dance (which he labels "curing dance" ). "Only the men dance
and enter trances," he tells us; but "the participation of the women is fun -
damental, for it is they who provide the musical framework which makes
the tr ance experience possible . T he men are quick to acknowledge this,
and say that the success of the dance is dependent on the perseverance and
sustained enthusiasm of the women." If R. B. Lee 's account is indeed ab -
solutely accurate, then the medicine dance could be seen as a shamanistic
seance in which the music is made not by bur for the shaman, a situation
that is in fact characteristic of possession, an d that , therefore, contradicts
the thesis advanc ed earlier on the relation of music co shaman ic trance . In
ocher words, the opposition I thought I had unveiled between shamanism
and possession with respect co "who makes the music" would be unfound -
ed. This point, therefore, warrants discussion.
At first glance, the two descriptions we have of the medicine dance, by
Lorna Marshall and R. B. Lee, seem to agree broadl y. Examined more
closel y, and from the viewpoint of their musical details, they contradict
one another. Lee (1968, 37 ) makes a very categorical distinction between
the women's and men 's role . "There is a fundamental asymmetry in the
roles the y play in the curing dance, " he writes. T he women sing, the men
dance. The women dance only occasionally, he notes in passing. "Only the
men dance and go into trance," he says without mentioning that the men
may sing. Lorna Marshall pro vides a much more derailed account. In -
deed, she writes that during the medicine dance, women play the principal
role in the singing and the men have an almost total monopol y over the
dancing, all of which corresponds co \\·hat Lee tell s us. But she also indi-
cates several times that the men (the "medicine men " ) also sing during the
dance: "The men dance and sing for some time (usu ally an hour or two, at
least ) before trances begin co come on," she writes ( 1969, 37 5 ). The rhyth-
mic stamping of the ir feet on the ground and the tinkling of small jangles
wrapped around their legs is added co their singing (ibid., 359 ). T his
acoustic dimension of their dance is so important from the musical point
of view that, when she discusses the role of various dance steps (ibid.,
364 ), Marshall concludes that the "p rimar y function" of these steps is nor
ro form choreographic figures bur to "produce sound." In other re spects,
she describes healing practices in the following terms: "The medicine
man, singing one of the medicine songs, couches the person , usually plac -
ing one hand on the person's back and one on his chest. In a moment or
two his singing changes co the formalized sounds called nl / hara" (ibid.,
142 CHAPTER THREE

370 ). These sounds emitted by the healer consist of groans, moans , high-
pitched cries, and sighs of all sores; they display a certain formal structure
because they are constantl y repeated , ha ve a rh ythmic organization , and
are sometimes slightl y sung. T hus the healers rake an act ive pare in music
m aki ng, both before the ir trance begins, in order to help induce it , and
also during thi s tr ance , in order to practice healing.
Thi s means chat , wi th respect to the rel at ions between trance and "the
one who m ak es the music ," the situation amo ng the Bushmen represents
an intermediate stare between possession , as we have obser ved it up to this
point, and sham anism as it is generally defined. Contrar y to what hap-
pens in the ca se of the possessee, the Bushm an healer takes part in the
songs char induce his trance, but contrar y to w h at happens in the case of
the sha man , the m ain responsibility for the singing falls not upon him , but
upon the chorus of women; char is to say, upon a group that never under
any circumstances goes into tr ance . Th ere are at least two underl ying re a-
sons fo r chis situat ion . First, the Bushm an healer, unlike the Siberian sha -
man, is nor a sol itary figure: as we have seen, at least fifty percent of all
male Bushmen are healers; shama nism am ong them is thus more a collec-
tive affai r than an indiv idua l one . Second, it is through dance th at the
Bushman healer atta ins the trance scare. This dance is collecti ve and re-
quires such a continuous flow of mu sic char it can only be pro vided by a
group of people specificall y assigned to chis t ask. Once the hea ler h as gone
into trance and exercises his shamanic healing power, however, it seems
char his own singing, and not that of th e chorus, takes precedence from
the patient's point of view. T his demonstrates that fo r both sha m an ism
and possession it is important to make a distinction between the different
stages of trance.
Novv char this point has been cleared up, examining the case of the
Bushmen also poses, and in a particularl y intere stin g way, the prob lem of
the therapeutic role of music and char of the rel ations of mu sic and dance
to trance .
\Y/eha ve seen chat among the Bushmen, we could to some extent distin -
guish between, on the one hand, pre venti ve medicine invol ving the who le
group and consisting in the curing dance, and on the other, curative medi-
cine practiced by the healer during pri vate sessions and without the use of
dance (ibid ., 369 ). We are thu s faced with two different practices . But
both use music, or rather curi ng songs - the songs mentioned earlier th at
are cha rged w ith a magical power. The great difference between the two
pract ices is that the first is indis solub ly linked to dance and trance , w here-
as the second does nor make use of d ance and is nor even necess arily as so-
ciated w ith trance. Moreo ver, whe rea s the power of healing seems to be
exclusive ly in the hands of me n , both men and wom en can sing the curing
songs, either in the group da nce or in the private session . Lorna Marshall,
desc ribing the healing of a sick ba by under dramatic circumstances (ibid .,
Music, Shaman ism, Mediumsh ip, Exo rcism 143
374 ) notes the following: "Inst antly the wo m en began t o sing a medicine
song in full vo ice (no tentative beginn ing ). Men from both gro up s began
to cure the baby. In minutes, two of the men from the pa rents' gro up and
one from the other group went abruptly into frenzied tra nce and soon fell
unconsc ious." She points out that in ocher serio u s cases, the med icin e men
she observed at work also went into trance, but only "into light tra nces"
(ibid., 373 ). With respect to two healing sessions in volv ing more benign
illnesses, on the contrary, she describes the medicine man as working
alone, beginning to sing without any preparation , making the nl / hara
sounds, rubbing the patient wi t h his forehead, flutter ing his hands ov er
his bod y. She adds char the medicine man did not show the slightest sign of
being entranced in either of these cases.
From all chis only one thing, amon g all chose with wh ich we have dealt ,
appears indispensible to the healing process in every case, and this is the
singing, under its double aspect of medicine songs, sung by men or women,
and of nl / hara sounds produced by the healer alone. What are these
songs like?' 5 For the sake of simplicity, let us sa y that aside from the
nl / hara sounds, they display the same musical characteristics as those
th at define the Bushmen 's vocal music in general: yodel ing, contrapuntal
polyphony, and vocalization (that is, a total, or almost total, absence of
words ). Lorna Marshall makes two extremely important observat ions
about these songs. On the one hand she writes: "Despite their believed su -
pernatur al origin and the fact that the y are charged with nl um, the songs
are no r set apart exclusi vely for the curing ritual. The y are sung every -
whe re by ever ybod y, more than any other music, to enliven casks and
miles of wa lking, to beguile hour s of leisure, and to delight and soothe ba -
bies" (ibid., 368 ). Interpreting information she has collected, she adds
chat it seems to her that "the nl um in medicine songs is thought to be inac -
tive until it reacts with the nl um in rhe medicine man " (ibid ., 368 - 69 ).
This interaction of the rwo powers would obviously nor occur under just
any circumstances, but only on the condition that the medicine man "actu -
ally be intending to cure and thus be seeking to evoke the power of the
nl um" (ibid., 369 ). Moreover, it should be added that according to Mar -
shall the dance steps also have nothing specific about them: "In the medi-
cine dance, the dancing is always the same whatever song is being sung,
. . . using the same basic step ... it is this step char a !Kung will use,
when spontaneously, on occasions quire unconnected with medicine
dance, sudden exuberance leads him to dance along for a moment or two"
(ibid., 3 63 ). This demonstrates that, as Roger Basride observes ( 19 5 5,
501 ) for possession music, amo ng the Bushmen, it is the "total situation"
char confers power to medic ine songs.
Unlike the medicine songs, which can be sung (and recei ved in dreams )
by either men or women, the nl / hara so unds are emitted only by the heal -
ers, that is to say, by men. These sounds are not necessarily linked to
144 CHAPTER THREE

trance or to dance, but the y are indissociable from the practice of healing
itself, and are thus in a sense its acous tic expression . I may also say in p ass -
ing, though without drawing the slightest conclusion, that one cannot ·
help but comp are these sounds, as described by Lorna Marshall (1969,
370 ), and as we can hear them on the recordings made by the Marshall ex -
pediri o ns,' 6 w ith those emitted by Amerindi an sham ans , which have been
recorded, in p articular, by H. Le Besner ais,' ' J. Huraulr,' ' and Audrey
Butt. 4 9
Viewed within the perspective of the relations betwee n trance, d ance,
and music, the preventive and curative aspects of Bushmen shaman ic heal-
ing practices differ from one another because each has a distinct configura -
tion of its components. This is made clear by table 3 .1, in wh ich those
components present in he aling are indicated by the sign + , those absent
by the sign - , and those that can be either absent or present by both signs
together ( ± ). Th e lette rs m and / indicate their distribution by sex. One
can see that the healing power of curing song s is nor nece ssaril y associated
w ith tr ance , any more th an w ith dancin g o r w ith the participation of
w omen. Earlier , w e saw th at a medicine so ng is not effecti ve unless it is
sun g by a he aler and w ith the specific int ention of healing. T he songs
themsel ves can posse ss either a great er o r a lesser degree o f magic al po w er.
Thu s w hereas o ne so ng m ay be able ro "cure any sickne ss" (ibid., 36 7 ), an-
orher m ay ha ve become wo rn our and lo st all o f its pow er (ibid., 368 ). In
ever y case, how ever , th e minim al condi tio ns th at must be present for a
son g ro h av e he aling p ow er are th o se we just menti o ned: the inherent
po, ver of the medicine song bec o mes effecti \·e onl y if it is in synerg y with
that of the healer. In other words, for the Bushmen a medicine song is not
therapeutic in and of itself but it is onl y one element o f a therap y. Let us file
this inform ati o n in the d oss ier o f rese arch abo ut musicorher ap y.
1 ow let us turn ro music from the view point o f its relations, not w ith
healing, bur with trance . Among the Bushmen, trance can occur ro ver y
var ying degrees, from "light " tr a nce ro extremel y violent trance resulting
in fainting and coll apse. From Lorna M arsha ll's description , we would
probabl y be justified in deducing that two condition must alwa ys be ful-

Table 3 .1 The Relations of Song, Dance, and T ranee to Healing Practices and Sex-
ual Differentiation

Prac rices

song n// ham dance tr ance

Healin g m f m f m f m f

preventi\'e + + + - + - + -
cu rari,·e + ± + - - - ± -
Music, Shamanism, Mediumship, Exorci sm 145
filled in order for violent trance to occur. On the one hand there must be
music (which is co say, essentially, singing ), and on rhe other there must be
either dancing ("medicine dance" ) or serious illness char in volves ipso facto
rhe medicine man in a dramatic situation. One might perhaps take the
view that the long preparation provided by the dance, which "wa rms up"
the healer 's power and brings it to a boil, creates an emotional state within
him that is equivalent, as far as disturbance of his consciousness is con-
cerned, to that resulting from the critical situation with which he is sud-
denl y faced when he must heal a seriously ill patient. Whatever the case,
although music appea rs to be indispensible to triggering trance, for all
that, it must still be combined with a certain emotional situation .
The situation that interests us here is the one that materializes whe n the
"medicine dance " cakes place. This dance lasts for several hours, some -
times for a whole night if nor longer. The warming-up period preceding
rhe first trances ma y cake an hour or two. During this rime, as we have
seen, the women sing and clap their hands whi le the men sing and dance,
loudly stamping-pounding, we could say - the ground, so that the little
rattles at t ached to their legs jingle to the rhythms of the steps. For Lorna
Marshall (ibid., 373 ), "rhe powers of autosuggesrion char induce trance"
are greatly fortified by "the loud singing char assails the ears for hours, the
exertion of dancing, the repetitiousness of the rhythms, the ph ysical near-
ness to ochers and the synchronization of movements with ochers." This
singing is wordless (which is a rule among the Bushmen, whe ther one is
dealing with healing songs or not ), contrary to what happens in posses-
sion. !rs action cannot therefore be attr ibuted to the emotional impact of
words, and is therefore purely musical. Marshall and Lee both stress the
passion with which the women sing, and also the loudness of both their
singing and their handclapping. The trance itself is associated, as we have
heard , \\·ith hear, fire, and boiling. 5° One is remin ded of Zempleni's obser-
vations about the wa rmin g-up music pro vided by griots to start off the
ndbp . Among the Bushmen, as among the Wolof, the role of music is to
arouse the dancers' ard or and raise their excitement. At the auditory level,
the louder it is , che greater its effect as a stimulus; at rhe motor level, the
more passionate it is, the more pressing its invit at ion to dance. As we saw
in the case of possession dance, music in the Bushmen shamanic dance ap -
pears to pla y an important role in a general program of overstimula rion .
Abo ve all, however, it appears to be the condition sine qua non for danc-
ing, and it is ulrimaceh· the d ancing that produces the decisive effects; the
face char only the dancers go into trance, w hile rhe seare d women do nor,
makes chis quire evident.
In ,,·hat does chis dance consist? Lorna Marshall describes it as ten se .
Th e body is "held erect o r leani ng forward bent at the hips," and the
movements of the torso are so controlled chat the dancers look "like scat-
146 CHAPTER THREE

ues being carried by the danc ing legs" (ibid., 363 ). As we saw, the men vio-
lentl y stamp the ground in an attemp t to make as loud a nois e as poss ible .
In the long run , does dancing in this manner entail a particular ly great ex~
pense of physical energy on the part of the dancers, o r does it incu r p ar-
ticularly significant metabolic disturbances? Th is is somet h ing we do not
know.
Let me sum up what we have learned from this visit to the Bus hmen.
Music fulfills several functions in the "medicine dance": ( 1) it incites the
men to dance; (2 ) it contributes to the entry into trance of those among
them who are healers; and (3 ) it leads to the practice of magical heal ing.
The greater part of the music is provided by the chorus of women, but the
men also sing. The chorus of women is the one that plays the essential role
in inciting dance. On the other hand, it seems th at the healer 's own sing ing
is of prime importance in the actual healing practices. Be that as it ma y, the
healers play an active role in the music, w hose purpose is to make them
dance and go into trance, w hich is contrary to what h ap pens in posses -
sion, yet at the same time the main burden of the music making rests on a
group that does not go into trance (the women ), whic h is contrar y to what
happens in shamanism . Viewed from the standpo int of the relations be -
tween trance and "who makes the music," the case of the Bushmen thus oc -
cupies an intermediate position between posses sion and shamanism, a fact
that weak ens , but does not invalidate, the oppositions dra w n in the pre -
vious chapter.
We should also remember that ( 1) music does not trigger trance unaided
and that it can work only in conjunction either with dancing or with an
emo tional situation; and (2 ) in the opinion of the Bushmen themsel ves the
efficac y of their "medicine so ngs'· certainl y resides in their magic power,
but it equall y resides in the power of the healer himself and his conscious
use of the song for healing pu rp oses .
Last, let us remember that, for Lorn a Marshall, the Bushmen's "medi -
cine dance" should be seen above all as a means of reasserting the unit y of
the group and reinforcing awa reness of that unity amo ng all of its mem -
bers. The wo rd s "u nited " an d "unit y" recur three times only several lines
apart in her conclusion (ibid., 380 ): "In their singing, clapping their
hands, and dancing, the people ar e united, ... people unite and dance
together whatever the state of their feelin gs ." E ven if the y hurl accusations
at one another in the course of the dance, "the next moment, the people
become a unit, singing, clapping, mo ving together." I shall return later to
the importance of this aspect of things.

As a sequel to wha t we have just learned abo ut the Bushmen, I would

like to ad d a few observations on tr ance among the Pygmies, their distant
neighbors. Both the Bushme n of the Kalah ari desert and the Pygmies of
Music, Shamanism, Mediumship, Exorcism 147
the great eq uaroria l fo rest are, as we know, nom ads w ho live by hunr in g
and gathering. Alth ough th ey are nor of the sa me "race," both are peoples
of ve ry small sra tur e . Although the y live th ou sands of m iles apa rt , with -
out any contact between them and in total ly diff erent environme nt s, their
music, or more precisel y their sing ing (since the y have no instru ment al
music, or very little ) is extraordinarily similar in some ways (Rouget
1957; Lomax 1968; Frisbie 1971 ). In certain respect s their dances are also
very similar. Among the Bushmen, as we have just seen, songs an d dances
are ro a gre at extent used to induce trance, and tranc e plays a role of great
importance in their lives . What about the Pygmies? In cont ras t to the
Bushmen, they are thought nor ro practice trance at all (Bourgu ignon
1968, 20 ). Colin Turnbull (1961 ) makes no mention of it, so we can be
sure that trance is unknown among the lturi Pygmies he studied. The same
is nor true, however, of the Sangha Pygm ies, or at least of tho se I was ab le
to observe in 1946, that is ro say the Babinga (Bangombe and Babenzele )
of the Ouesso region. One day during a dance, the Bangombe hunte r who
was leading a file of dancers suddenly became entranced before our very
eyes (exhibiting shaking, loss of balance, abundant perspiration, bewil -
derment and then collapse ). A little girl then brought cool wa ter, which
was applied to his face as he lay unconscious on the ground, a situation
comparable to what Lorna Marshall (1969, 378 ) describes for the Bush-
men. The cold water brought the hunter grad uall y back to his senses, but
he remained sitting there for a long while, obviously exhausted, wh ile th e
orhers we nt on dancing w it hout him . I can report nothing more abo ut this
trance, except th at the group obviously perceived it as perfectly no rm al.
.\tfy stay among the Pygmies was too short for me ro discover anything
more on the subject.
Let us simply note that, contrary to what has been said, trance, in rela-
tionship w it h dance and song, does occur among certain Pygmy gro ups, a
fact chat is of some interest for mapping the distribution of trance and for
comparative stud ies wi th the Bushmen.

EXORCISM'

The words possession and exorcism are so frequently linked together in

ethnographic literature and in Western rel igious th ought - and the assoc i-
ation is clearly not fortuitous - chat one might almost say rhe one inev ita -
bly trails rhe other like a shadow. Both are often seen as covering two nor -
mally complementary aspects of one and the same phenomenon, or,
rather, the one is seen as the question and the second as the answe r that the
first raises. Given this basic approach - wh ich is evidently not ours - one
could define the va riou s types of possession in terms of the greater or lesser
importance given to exorcism in the response made to them. If possession
148 CHAPTER THREE

we re represented as a circle, the position of exo rcism cou ld be dep icted as

going from zero, expressed as a totally empty circle and conveying the to -
tal absence of exorcism, to a full 360-degree turn, cove ring the entire sur - ·
face.
In th is latter case, in which exorcism is the only poss ible re spo nse to
possession, we would be dealing with possessions of the Loudun type,
which ultimately are rejected by society or, as we labeled them, rep roba te.
They do not interest us here since they do not give rise to any form of
cult .; , All the same, we ought to note that the Loudun possess ions, wh ich
occurred outside of the official religion, cannot be reduced - an y more
th an those concerning us here, which are completely integ r ated into rel i-
gious life - to cases or a series of cases of individual hysteria. Th ey ha ve so
many stereotypiced feature - use of a secret language, numbe r and hier-
archy of demons corresponding to the personality of the possessed , to
mention bur a few - th at they certainly should be connected wi th older
possession cults, of which the evidence sugge sts they are (partiall y at least )
a survival or resurgence. However this mav be, possession of the Loudun
tvpe is not accompanied by music of any kind . As \\·e have seen, the princi -
pal function of possession music is to establish communication with the
gods . There is therefore no reaso n to use music when the possess ion, far
from being desired, is rejected. T he events at Loudun therefo re confirm
the logic of our theor y. We \\·ill linger on them no longer.
In the first ca e, on the contrary, in \\·hich possession is unreservedly
\\·elcomed br society, we are dealing with such cults as the vodun in D aho -
me~-,5' which give rise to possession seances totally devoid of exorcism . 53
This does not mean that exorcism prac ti ces are completely u nkn ow n in
vodun \\·orship. When someone has been struck by ligh min g, the high
priest of Khevioso, god of lightning, comes to exorcise the victim's
house.;, But such exorcism rituals are one thing and ceremonies in wh ich
possessions occur quite another. Each has rather different orientations
and, needless to sav, onh· the second kind inte rests us. In Vietnam, hdu-
b6ng ceremonies are devoted exclusiveh· to the "gua rdi a n Spirits, Genii,
and, in an accessory \\'a\' . the Souls of the Dead'·; the "evil Spirits . . . do
nor intervene in am· \\-a\' '' (Simon and Simon -Barouh 1973, 76 - 77 ). Be-
rn·een possession of the Loudun kind and the one found in vodun worship
or rhe hcii1-b6ng ceremonies. bern·een possessions that are rejected on rhe
one hand and possessions rhar are entireh· \\·elcome on rhe o ther, there is
room for a \\·hole series of intermediate cases.

Possession and Pact -Making : African Examples

,-\m ong the VaNdau of Mozambique rhe possessed are characterized b\'
rn·o rypes of possession rhar coexist and are complementary (Jun od 1934 ).
Music, Shama nism, Mediumship, Exorcism 149
T hose of th e fir st type are bo u nd to be exorcised. T he functio n o f those of
the second typ e is to exo rcise th ose o f the fir st. 55 W hethe r the y belong to
one or t he other of t hese two catego ries, t he possessed are all occ upied by
spir its of the Ndau ancesto rs, but the vas e major ity are possess ed by m a-
levolent spirits, whe reas the remaining few are poSSf'S ed by guardian spir -
its . The first undergo their possession passivelv since they depend ent ire ly
upon the seco nd ca tegory to rid them of it. Contrary to the first, the latt er ,
called nyamusoro, undergo initiation, at t he end of which they are ab le to
fulf ill their funct ion - that is, to rid the ochers of their possessio n . J unod
calls them exorc ists . Unlike the passive majo rit y of possessed pe r so n s,
whose sta te is caused by spirits ",,·ho enter into living beings and to rm ent
them unt il they do all that is demanded of them" (ibid., 273 ), t he "exo r-
cis t" de r ives hi s "sp iri tual powe r " from "t he sp irit of a dead per so n w ho
has possessed him, not to harm bur to pro tect him" (ibid., 283 ).
As descr ibed b y Junod (ib id., 286 ), the exorc ism seance (dur ing w hich ,
at the 1equest of the passively possessed person, che nyamusoro rids h im
of the tormenting sp irit ) appears almost more like a shamanic seance -
albeit w ithout any journe y- than a possession seance. I sha ll not enter
into further detail on this aspect of things here . We ,,·ill, howeve r , rake
nore of the method the "exorcist" uses to liberate his pat ient (ibid., 2 87 ).
Junod describes chis method as it relates to d iverse "categories of ances tr al
spirits," the majori ty of whom are Ndau . It q r ies according ro the ca te-
gory and J unod's informa tion is more or less detailed fro m case ro cas e . In
order to be appeased, each of the twelve sores of spirits mentioned re-
quires offe rings and sacrifices, but some are more ·'difficult to sa tisfy"
th an othe r s. T he re are those who "deign to appear only when ever yo ne is
drunk" (ibid., 277 ). One can make ome spirits "rejoice " by p layi ng the
flute for them (ibid . , 275 ), ochers by "handclapping and dancing" (ib id.,
28 0 ). The spirits of hunters are "propitiated" by imit at ing the ace of hunt -
ing. To "please" certain female spirits . o ne dons animal skins ,,·ith fringes
and pearls (ibid., 280 ). Clearh·, these are ,-er,· fr iendh· ,,·ays of exorc izing .
.-\nd indeed, although the seance described by Junod does inrnlve '·liberat -
ing" a pos essed person of the spirit ·'tormenting " him, it is far from being
true exorcism . The spirit in que stion is nor driven our; it is appeased,
made to rejoice, provided ,,·irh some kind of satisfaction, and ,chen polite -
h· escorted back inro the bush: "There is your place, yo ur village" (ibid.,
2 7 ). There is a certain contradiction in terms invoh·ed, in fact, \\"hen one
speaks, as does Junod , of "the Ndau exorcises' propitiation ceremonies"
(ibid., 2 7 3 ): either one propit iates or one exorcises; one canno t do bot h.
Also in sou t hern Africa, the T honga, neighbors of the VaNdau were de-
cribed by H . A . Junod (father of H . P. Junod ) as responding to poss es-
sion with exorcism; possession among them is attributed to foreign (Z u lu
or jao ) and consequently malevolent spiri t s fro m whom one must be r e-
150 CHAPTER THREE

lieved . Us in g H. A. Jun od's ow n desc ription, Bastide has shown (1 972,

103 - 4 ) that the evil spirit in real ity reconciles himself with his victim and .
becomes his gua rdi an sp irit. I sho uld add th at the real purpose of the
"strange and terrible exorcism" (Junod 1913, 440 ), wh ich in vo lves an "in-
fernal" or "h ideous" din of drums sometimes for days on end, is no t to
drive the spirit out at all, but to force it "to revea l itself, to decl a re its
name" (ibid . , 441,443 ). On this particul a r occasion "the patient will sing
his song, ... a song which will be henceforth his" 56 (ibid . , 445 ), and
which will serve in the future, at th e appropriate moment, to induce fur -
ther tr ances in him. In this case, we are clearl y at the oppos ite pole of exo r-
cism and reprobate posse ssion.
In the Central African Republic among the Ngbaka -Mandja (Arom and
T aurelle 1968 ), possession wo rk s in the sam e way as amo ng the Va d au
in the sense that there are two categories of possessed: the he alers, who go
int o possession trances in ord er to he al , and the so -to-speak occasion al
possessed person, who comes to the healers ro be cured . As among the
VaNdau, the healer is simila r in many respects to the shama n- he is the
o ne who induces tr ance in the sick person, sings the appropriate so ng s
while shaking his rattles, and directs the musicians . Hi s songs clearl y
show th e ambi val ence of the re lat ion s with the deities re sponsible for the
p oss ession and the illne ss . On the one hand he calls to them, "Come and
help me cure the sickness!" an d o n th e oth er he threatens them: "If yo u tr y
to come back into her and poiso n her again, yo u shall all die! " At the end
of the ceremony, after the offe rin gs have been made , the healer add res ses
the divinities again : "T his is for you ... eat your sha re . . . do not trou-
ble her any more " (ibid., 102, 103, 104 ). Her e, as amo ng the VaNdau, oc -
casion al possession results in a relation \,·ith the divinit y th at could be
termed pe aceful coexistence : let both sides stay in their own place.
In the Haus a bori, which exhibits aspects of possess ion that are very dif-
ferent from the foregoing, initiation m ay be broken down into two succes -
sive phase s, which J ac queline Monfouga-Nicolas (197 2, 11 0ff. ) describes
as th e "exo rcism" pha se and the "imposition " phase. During the first , that
of exorcism, the jakeso -t he officiant conducting the initiation - pla ys a
similar role, all things being equal, ro that of the nyamusoro, the Tdau
"exorcist ." Thi s first phase lasts for chm· nights (ibid . , illus. p. 112 ). Its
purpose is, as it we re, to w ipe the slate clean so that the benefice nt di vinit y
can t hen come and take possession of its "ma re." This beneficent divinit y
can belong ro the category of either white or dark gods ; the latter (kan -
kama ) can cross over into the category of wh ite gods if it be comes benefi -
cent for the ne oph yte. There are , howe ver, certa in kankama who will al-
ways remain uny ieldingly dark . During the first nm nights, the kankama
manifest themsel ves when they are summoned by mus ical instrum ents .
For each one, the initiating jake so goes into a light trance (ibid., 116 ), an d
Music, Shamanism, Mediumship, Exorcism 151
the neoph yte being initiated "seems absent." On the third night, however,
the "struggle with the dark gods" takes place (ibid., 120 ). They arrive in a
cohort, as on the previous t-"IO nights, but this time "eac h of the gods in-
voked will swoop down upon" the neophyte. "After each invocation the
god refuses to leave the initiate's bod y," and she muse then "fight with him"
to make him leave her. At the end of this struggle, which lases for several
hours , one of the white gods, who is assigned "the task of driving our the
kankama for good" (ibid. , 121 ), comes and incorporates the jakeso. The
latter then struggles with the neoph yte she is initi ating, who is still pos-
sessed by the dark gods, and finally emerges "victorious from this strug -
gle ." As with the VaNdau, then, exorcism invol ves an exorcist who is pos -
sessed by a benevolent god and uses this power to drive out a malevolent
god from the body, not of a patient this time, but of a neophyte undergo-
ing initiation. The methods used, however, do nor appear to be at all the
same. There is no question of "satisfying" the kankama; they are dri ven
out. This is, therefore, a true form of exorcism, even though, during the
rwo nights preceding the struggle with the dark gods, the jakeso went
"into a light trance at the sound of the motto " upon the arrival of each
dark god (ibid., 116 ), wh ich in a way implies p ay ing homage to or at lease
acknowledg ing the kankama. Be th at as it may, the important thing about
chis type of exorcism is chat it constitutes no more than rhe seco nd part of
a process in two stages, since before the dark gods are expelled the y are
first summoned. Another import ant point ro remember is chat the sum -
mons ro rhe dark gods does nor differ in any way from the one to the wh ite
gods. Both are issued by musical instrum ents on w hich divinities' mottoes
are played.
Let us now move to Chad. In her article entirlrd "Musical Exorcism
among the Koroko," Moniqlle Brandil y describes a ceremony organized
for the purpose of exorcizing a sick woman ''because she had fallen prey ro
a genius" (1967, 38 ). After a considerable length oftime spent in organiz -
ing various participants - musicians, \\·omen singers, members (women )
of the college of the possessed, spectators - the seance begins with the pos -
session of the sick woman. Th e musicians pla y music addressed co the ge-
nius inhabiting her. The sick wo man dances and then a number of ocher
women also become possessed and dance . The seance ends when the last
woman has been "freed": "All the women have now been freed," the au -
thor notes (ibid., 49 ). The woman singe r, who played the central role in
the sea'lce, then embodies the "ma r about" genius and in this capacity begs
the "ch ief " of the genii to order to all his subjects "to leave the wo men in
peace" (ibid., 50 ). These data lead Monique Brandily to comment on the
ambiguity of the ceremony, which, she observes, "cons ists at the same
time in driving our the genius in order to obtain a cure," and in provoking
"several cases of possession among the women present" (ibid., 51 ). What
 152 CH APT ER THREE

is its real aim? Is it to drive out the genii - or as she prefers to call them, the
demons (ib id., 38 n.1 )? Or is it to make them come? To use Luc de
Heusch's terminolog y ( 1971, 229ff. ), is it exorcism or ado rcism? ·
This question, which is posed by so man y possession cults, is by no
me ans new. In his work, La possession et ses aspects theatraux chez Les
Ethiopiens de Gondar, Michel Leiris (1958, 34 n.2 ) makes the following
observation: "It is improper co apply the term 'exorcism' to the practices of
the ziir brotherhoods; their goal is in fact co make a pact w ith the spirit
rather than to expel it. " This observation can in fact be applied to all cults
in which the rites are directed toward expelling the spirit responsible fo r
possession. All available descriptions demonstrate that this expulsion is
only obtained if a pact with the spirit has been made beforehand . Before
the genii "leave the women in peace," as ?vionique Brandily puts it, the
women must first publicly identif y themsel ves with them, for by reproduc-
ing the geni i's outward behavior in dances that imitate them, the women,
in fact, become identified with them. But this identification, this posses -
sion, cannot take place unless one has already summoned these genii, and
this precisely is the role of the musici ans and women singers . The ceremo -
ny's primary aim, therefore, is to summon the genii, not co drive them
away.
To return to the healing seance described b y Monique Brandily, from
the sick woman's point of view, events unfold - whethe r consciously or
unconsciously is of little import ance - someth ing like this: I am sick be-
cause a genius is rormenting me, I become that genius; satisfied because he
has possessed me, he ceases to rorment me , I am set free . Is this exorcism?
Broadly speaki ng , it could be, since ultimately the sick wo man has ex o r-
cised or chased out her illness. Strictly speak ing, it certainly is not. We
speak of exorcism when the divinity who comes to be embod ied is seen as
a demon. But there is every reason co believe this is the case only because
among the Kotoko, who live in a region com·ened co Islam, there is a more
or less oven conflict between Islam and the old, traditional religion . Al-
though they once were deities - and still are, incontestably, in man y pans
of Africa - these genii have become demons. Thus people succumb to
them because belief in their power is not yet extinct, but the y then expel
them - at least this is what they say - because this is wha t the official or
dominant reli gion requir es. The trances of the female Tuni sian possessees
described by S. Ferchiou (1972, 60 ) take place with in a similar religious
framework.
In other pl aces, as the example of the Thonga demonstrated, the con -
flict manifests itself differently : there, the spir its that possess the patient
ar e foreigner and therefore enemies. Again the patient must get rid of
them, they must be driven ou t. But is this exorcism? Here again the same
process is used: in order to expel them, one must first enter into communi -

u...
Music, Shamanism, Mediumship, Exorcism 153
cat ion with them and appease them; in sho rt , one mus t hav e established
bonds with them, if not through possess ion, at leas t by m aki ng a p act.
What we need to remember from all th is is tha t in a great many cult s , in
black Africa at least, possession seances a re devoid of any elem ent of ex or -
cism. Whe n eve r t here is genu ine exorc ism t he ex p uls io n of the ge niu s,
however dramatic it ma y be, is me rely t he culmination of a process in
which integ r ation of the deity is a prerequisite and a sine qua non . Fin a lly,
the term exorcism is frequentl y used pejoratively, mainly because contac t
with another religion ha s led certain people to refer ro spirits as de m ons
when, in fact, for the adepts there is not hing demoniacal about them.
We must be ca reful not ro oversimplify, hov-1ever. Among the Mundang
of Chad (Adler and Zempleni 1972, 32 - 35 ) rhe enrire svsrem of posses -
sion consists, from the sick man's stand p oint, in being possessed by age -
nius ,,·ho is in fact seen as an "agent of sickness. " In this case, should we
compare the possession to a k ind of hom eop athic cure rhar treats diso rde r
with disorder , illness ,,·i th illness, or, as Firth puts it (1 969 , xi v), "f ights
fire ,,·ith fire "? Howe ve r th at m ay be, here again there is identification
,,·ith the divinity, initiation int o his cult, consec r ation of altars, and pe r-
forma nc e of sac rifice s . This is the imp o rt ant point for us.

Let us sum up. Possession has ofte n been sa id to require exorci sm as a

response for reaso n s that ,·ary from case ro case. Sometimes it is because
the descriptions a re no more than a projection of the \\Titer's own ideas
onto the facts observed. More often still, it is because the cult in question,
,,·h ich in the regi on ,,·here it is found rep re sent s either the autochthonous
religion or an ea rlier stage of a religion, is in conflict \.Vith a monothe ist ic
religion brought into the region more recentl v a nd oc cup ying a politicall y
o r morally predominant position. The spirits, the genii responsible fo r
possession are then seen by followers of the official religion as demo ns ro
be exorcised, an attitude that is frequently disturbing for the adepts.
The e latter then accept the fiction of exorcism all the more readily be -
cause the genii a re indeed tormenting them, because by expelling them o r
bY pretendi ng ro do so, one keeps everyone ha ppv. and because, in addi -
tion, th e essential as pect of their cult, which is to practice possession, is
safegua rded. There also are cases in "·hich possession is due ro fo reig n
pirits, a situation ,,·hich generally re su lts from recent historical ,·ic issi-
tudes an d consequentlv reflects a conflic t sit ua ti on, but of anothe r type . In
thi s case expulsion can take precedence, appa rentl y, a t least, and in cer -
ta in pares of the ritual, over the making of pacts . onetheles?, in orde r ro
expel these spirits, one must first embody them, which usu a lly entails a
preliminan- initiation stage. 5 '
I hope the reader will excuse me for having spent so much time on the
quest ion of exorcism but it is in fact crucial. If, as I believe, the relations of
154 CHAPTER THREE

m u sic ro possession are of the same order, give or take a few variants,
whatever the possession cult, it is because all these cults, notwi th standi ng .
their di versit y, exh ibit one and the same type of religious behavior. If they
did not, m y hypothesis wou ld then be unthinkable. It was therefore im -
portant ro establish that side by side with possession cults in wh ich the in-
trusion of the divinity is not onl y welcomed but even ardentl y desired,
there also exist others in which possession only calls for exorcism . If this
were so, the inner beha vior of the possessed person would necessarily be
different in the two case s. Indeed, logicall y it should be diametrically op-
posite. \Y/ehave seen that one of the main functions of music is to establish
communication with the gods and thus to create a situation of identific a-
tion and a lliance chat favors the kindling of possession. If relations wi th
the gods were conflictual, then the music would either have ro play a dif -
ferent role or be of a radicall y different kind. App arentl y, this is not the
case. This is why I was led to examine exorcism so closely, and ro deter-
mine whether it does indeed play the preponderant role, wh ich has often
been attributed ro it, in possession . le does not. Or if it does, the posses -
sion concerned belongs ro the reprobate (as at Loudun ) and not ro the cul -
tiva ted kind-in ocher words, it is not a possession cult - or else it is obses -
sion (in the medie val sense of the wo rd as it was described earlier ) and not
possession in the sense we ha ve agreed upon in chis book.
One could almost say chat the difference between possession and exor-
cism lies in the fact chat possession necessarily calls upon music at some
moment or ocher while exorcism, on the contrar y, does not make use of it.
This is fully illustrated by the situation Scheherazade Qassim Hassan de -
scrib es in Iraq. In her ch ap ter de voted to "the possess ion ceremonies of
black population" ( 197 5, 172ff. ) she distinguishes between "non -musical
ceremonies" devoted ro "exorci zing the spirits of Evil, " and "musical cere -
monies" whose function is "the invocation of beneficent spirits. "

Saul and David

The well -known story of D avid pla ying the lyre 5 8 ro calm Saul's delirium is
often considered to be the most well-known example of musical exorcism .
le therefore merits in vestig ation.
Lee us begin by looking at the va rious interpretations chat have been
made of it . \Y/hen he writes, with reference to the role of music in Israel,
chat "music drove the demons ouc of Saul's soul when Da vid played for
him," Cure Sac hs (1 940, 105 ) clearly interprets it in terms of exorcism.
The same is true of E. Dhorme (1956, 868 ), for who m David and his mu-
sic constitute the "remedy for the possessinn of which Saul is victim. " The
same is also true of Combarieu (1909, 86 ), for whom Saul is a possessed
person upon whom Da vid 's music exercises a "beneficial" effect. But when
Music, Shamanism, Mediumship, Exorci sm 155
he discusses the episode in which Saul, blinded wirh jealousy, attempts to
kill Da vid who is actually playing for him, Combarieu adds that when
music "in the spirit of possessed people is associa ted w ith painful memo-
ries," it is also capable of pro vok ing "terrible crises." For him, Saul is "a
well-known example" of this "do uble role of music ," which sometimes
consists in "creating a pathological state that is inherent in the musical
emo ti on itself," but which, on the contrary, in different circumstances,
"a rr ests his fury and restores the subject to a normal state." H. G. Farmer
(1926, 12 ), who, like Sachs, pays attention only to the beneficial aspect of
David 's music, notes that "the evil spi rit " who was affl icting Saul is con-
jured b y David 's kinnor . Here again we are back with exorcism . Unlike
the pre vious two writers, however, he does not spe ak of possession bur
emphasizes the magical power attributed to David's music in both Hebraic
and Arabic literature. Fin ally, A. P. Merriam (1964 , 111 ), who like Farmer
restricts himself to the calming effect of D avid's lyre on Saul, sees it as an
example of healing music th at should be classified under the broad head -
ing of music's physical effects, particularly musicotherapy. What are we to
think of all this?
To clarify things somewhat, let us begin by formulating the facts in the
terms we have adopted for this book. If the m atte r is complex, this is be -
cause Saul himself has a complicated relation to his trance. He has fre -
quentl y been described as a possessed person. Barton (1918, 135 ), for in-
stance, says so in his article on Semitic and Christian possession in the
Enc yclopaedia of R eligions and Ethic s. In my view, Saul's trance has two
aspects which must be clearl y distinguished , and neither of the two consti-
tutes possession in the strict sense of the word as I am using ir. On the one
hand he is tormented by "an evil spirit sent from God." In no case, however,
does he identify himself with th at spirit. In our terms, therefore, it is not
possession chat is in volve d bur obsession. On the other hand - and chis is
so mething chat has not yet been menti oned-he is also a prophet, a nabi,
and therefore suscept ibl e to being seized by prophetic trances, when the
sp irit of Jehovah descends upon him. Here again, ho\\·ever, there is no
ques tion of identific ati o n, sin ce chis \\·ould be unthinkable. Fo r us then,
Saul the prophet is inspired, no r possessed. Vie\\·ed in terms of its rwo as-
pec ts, what is the rel ati on of his trance to music?
We know rhe circum ranee in \\·hich Saul prophesied- and therefore
went into tr anc e-for the first rime. After anointing Saul as ruler of hi s
people, Samuel cells him, among other things: "and furthermore when
yo u come there to rhe cit y, yo u will meet a band of prop hers coming down
from the high place wi th a lyre, a tambourine, a flute, and a har p ' 9 before
chem; and the y will be prophe sying ecsta tically. Then the spirit of the Lord
shall suddenly seize upon yo u, and yo u shall prophes y ecstatically w ith
them and you shall be changed into ano th er man" (1 Samuel 10: 5-6 ). So
156 CHAPTER THREE

ir is while lisrening ro music rhar Saul goes inro his firsr rrance. And we
should nore char among rhe va rious insrrumenrs listed, rhose regarded by .
rhe Hebrews as specific ro rrance are rhe srringed insrrumenrs, nor rhe ca-
brecs or the flutes. Thi s is made clear by the srory of Elisha. Before he
prophesies at the request of the three kings of Isr ael, Judah , and Edom,
and before he performs a miracle to save their threatened arm y, the prophet
demands that a musician be brought to him. And "when the musician
pla yed music 60 the hand of the Lord came upon him" (2 Kings 3:15 ). The
Hebrew text does not indicare wha t insrrument the musician played,
whether harp or lyre, but rhe verb translated here as "played" in fact
means "playing with one's hands," and refers to a strin g instrument. 61 In
the inrroduction ro his edition of rhe Bible (19 56, xvii i), E . Dhorme stresses
the importance of this passage and talks about Elisha's "t rance ." The
prophet 's entr y inro rr ance thus must be atrribu ted ro thi s string music .
This being granted, we can now turn to the texr that relares the circum-
stances under which David was summoned ro pla y before Saul: "Now the
spirit of the Lord had departed from Saul and an e\·il spiri r from the Lord
terrified him . So rhe servants of Saul said ro him, 'See now, an evil spirit
from the Lord is rerrorizing you. Ler now your sen ·ants who are before
you speak: let rhem seek for our Lord a man skillful in playing the lyre.
Then whenever the evil spir ir comes upon you he will play wirh his hand,
and you will be well'" ( 1 Samuel 16:14 - 17 ). Saul demands rhar a good
musician be found, David is broughr ro him, and Saul takes him into his
se rvice. And verse 23 reads: "So whenever the e\ ·il spirit from God came
upon Saul, Da vid would cake the lyre and play with his hand, and Saul
would be relieved and feel resrored and the evil spirit wou ld depart from
him." Now, if we re stricted ourselves ro rhis text alone we wo uld have ev-
ery reason to chink, as Sac hs does, chat by "playing with his hand" David
dri ves out the evil spirit, or in oche r words exorcizes it. Bur in chis case we
are faced wirh a contradicrion. Earlier, ir was by "playing wi rh his hand"
rhar rhe musician induced Elish a's propheric rrance; now ir is also by
"p laying wirh his hand" rhar Da vid succeeds in calming rhar of Saul. How
can rhis contradiction be res oh·ed)
Ler u go back to the beginning of rhe quorarion from Samuel. "Bur rhe
Spirir of che Lord departed from Saul, a nd an e\·il spirir from rhe Lord
rroubled him. ,. Elsewhere ( l Samuel 18: 12 ), \\·hen Saul, raving under rhe
influence of a n evil spirir, rries ro kill David , we read once more rhar God
"had deparred" from him. Thi s wirhdra\val of rhe Spirir o f rhe Lo rd rhus
appears ro be a precondiri on for rhe arrival of rhe "evil spirit from the
Lord. " (And one cannot help thinking here of rhe exchange of souls we
mentioned several rimes in African possession cases. ) This being so, why
should we not suppose chat if rhe evil spirit deparrs from Saul when D avid
"plays with his hand" before h im, it is less because rhe music drives him
o ut than because ir brings back in Sa ul rhe Spirit of God rhat had left him?
Music, Shamanism, Mediumship , Exorcism 157
We can then say that by playing his lyre Da vid reconciles Saul with the
Spirit of God, a situation that is no longer in contradiction with that of the
musician using his stringed instrument to induce Elisha's trance, a trance
that is inspired by the spirit of Jehovah . We are now in line with the logic
of the relations between music and possession as it has gradually eme rged
in the previous pages. It is true, as I have said, that neither Sau l nor Elisha
is possessed in the strict sense of the term as I am using it: they never identi -
fy themselves wi th God. Instead, they are inspired, or prophets (nabi ). Ad -
mittedly, the dividing line between inspiration and possession is often
rather thin. In this inst ance, see n from the standpoint of their relations to
music, one might even claim that it has d isappeared altogether .
By interpreting David 's music as having the effect of inducing God to be-
come present again and thus driving out the evil spirit as an aftereffect, we
remain within th e general logic of a well -attested system . By interpreting it
as exorcism music, not only are we faced \.Vith the contradiction al read y
mentioned, but we also run up against yet another difficult y.
What is generally considered to be exo rcis m music seems to act in only
one of two ways, either by virtue of noi se or by the magic of incantations.
Here there is no question of either being involved. Like the Hebre ws, the
Babylonians thought that it was necessary for the "good" spirits ro have
left someone before he or she could be possessed by "demons" (Barron
1918, 133, 135 ). Surviving Babylonian inc antatory texts for use in exor-
cising evil spirits express this belief very clearly. We thus have good reason
to think that if it was a matter of driving an evil spirit our of Saul, whose
·'good sp irit " had left, David would have chanted incantations. No such
thing is mentioned in the Bible.
If my hvpothesis - that Da vid used his lvre to effect a reconciliation be -
rn·een Saul and God - seems unjustifiable, there is really onh· one othe r
that could be acceptable; that is, Farmer·s . He depicts the lyre as a magical
instrument and Da\·id as a musician famou s fo r his magical power over
beings and things. This places him in the ear -Eastern tradition, which
according to Farmer originated in Mesopotamia, that attributes powers
to mu ic derived from complex symbolic or numerical relations with the
cosmos. We shall encounter this tradition again in the chapter on the
Arabs. These powers, \\·hich we re magical in nature, formed the basis of a
theory of healing by means of music. Here we return to Merriam's theor y
of musicotherapy, which in fact is perfecrly compatible with Farmer's
ideas.
Th e contradiction \\·e spoke of earlier could then be resolved in this
\\.a\·; \\·hen the musical instrument triggers the prophetic trance in the nabi
(it does not matter whether it is Elisha or Saul ), the mechanism at work is
perhaps not exacrly that found in possession bur it is nevertheless very
close to it; when D av id's lyre calms Saul, on the other hand, another musi-
cal tradition is at work; that of magical music.
158 CHAPTER THREE

The case of D av id pla ying the lyre ro calm Saul's frenzy can thus be in-
terpreted in three different ways. First, it can be seen as a case of musical
exorcism. I h av e al re ady stated the reasons for rejecting this hypothes is.
Second, it can be seen as an exa mple of mu sicotherapy. But this means
that we would resort tO a notion (musical m ag ic) a bsent from the text, and
that we wo ul d ignore an essential situational fact, wh ich is the reference ro
intrusion by an "ev il spirit." Third , one can see- as I do - Saul's obsession
by chis evil spirit as being in so me way the re verse of God's absence. By
pla ying hi s lyre D avid reest ablis he s God 's presence, wh ich means char he
resrores in Saul , in an atte nu ated form, the scare of inspired prophet char
he had momentarily lose. Granted, chis is a rather complica ted interpreta -
ti o n, but it does have the advantage of raking into acco unt all th e textual
data. It also rem ains within the gene r al system of relations between music
and prophetic inspiration among the H ebre"·s, a system that is in fac t
qu ire close to the even more gene ral one governing the relations between
mus ic and possession .
One fin al point concern ing Saul 's furor, in his attempt to transfi x David
\\·irh a ja\·elin while David is in the very process of playing his lyre for him.
Combar ieu interprets the prophet's homic idal delirium as a possession fir
brought on by the music. 1o rhing in the passage from Samuel jus tifi es this
interpretation . After it rec ounts D av id 's military success es and Saul 's rea-
sons for rak ing umbrage \\·irh chem, the text reads: "On the next da y the
e\·il spiri t from God seized up o n Saul, and he was filled w it h prophetic
frenzy within the house, while Da vid ,vas pl aying wi th h is hand as he did
each day. Now Saul had his spea r in his hand; and Saul lifted up his spear,
saving, 'I will pin D avid ro the wall. ' Bur D avid escaped from his presence
rn·ice. Saul wa afraid of Da vid because th e Lord was with him and had
departed from Saul" ( 1 Samuel 18 :10- 12).
Contrary ro what is clearly stared in the rwo ocher passages quoted from
Samuel char deal, first, with the prophetic tr ance char seizes Sau l for the
first rime while he is in the compan y of other prophets , and second, w ith
his state of obsession calmed by D avi d's lyre (passages char both associate
music directl y wi th tr ance, first ro induce it; then to end it ), this third pas -
sage does not in any way impl y chat th e sound of Da vid 's lyre is responsi -
ble for Saul's furor. An evil sp irit from God (read : jealousy ) co mes upon
Saul and provokes his fury. Ir so happens char at chis po int David is playing
his instrument , as he does every day. The text does nor connect the two
fac es. It is clear tha t Saul 's murderous fury mu st be attributed to th e sight
of Da vid as a ri \·al warrior, not ro his presence as a musician .

THE CASE OF T AR..\NTIS:\!

As \\·e know, rarantism has been long rega rded as a particularly character-
istic and spectacular form of musicorherapy . Opin ions on chis subject
Music, Shamanism, Mediumship, Exorci sm 159
have had to be considerably revised as a result of the work done by Ma riu s
Schneider (1948 ) and Ernesto De Martino ( 1961 ). The interpretations ad-
vanced by these two scholars are very different. Both, to be sure, acknow l-
edge that musicorherap y is an important factor . Bur whereas Sch ne ide r
sees it as simpl y one element in a vast system of symbolic representat ions
one might call as trologic al in nature, De Martino sees it as a form of exor -
cism functioning w ithin the psychoanalytic logic of a religion based on re -
morse. Borh 62 also associate rarantism as a religious phenomenon wi th
possession cults as a whole, bur the y do so with what one might term re -
pugnance, so that the y present this aspect of things as entirely secon d ary .
1eirher the ·word "possess ion " nor th e word "trance" ap pe ars in the index

of either book .
Is rarantism pos ession) Musicorherap y ? Exorcism) We return, though
in a ,,·holly different context, to the questions we just posed in the case of
David and Saul. What is the answer? Here again the exorcism theory has
to be discarded, as does that of musicorherapy. Examined close ly, raran-
rism turns our to be nothing more than a particular form of possession . Ar
least, this is rhe hypothesis underlying everything I ha ve written in preced -
ing chapters on rarantism 's relation ro music . Ir is now important to dem -
onstrate this hypothesis, as I hope to do in rhe ne xt few pages. Indeed, it is
particularly important for my present purposes ro clarify rhe relation be-
rn·een rarantism and music, since it represents the f->esr-known example of
trance in all of Europe. 63 We have rel atively early descriptions of raran-
tism, and it has been the object of recent investigations. It is, moreover a
ritu al in which music and dance pl ay an essential role .
Let me begin ,,·irh broad summa ries of the two theories already men-
tioned. In his La danz a de espadas y la tarentela .. !arius Schneide r looks
at onh- one aspect of rarantism-admittedly a particularh· interesting one
fo r musicologists - that of "medicinal"' rites expressing a cosmological vi-
sion of the ,,·orld characteristic of megalithic ci,·ilizations, or ar least of
those civilizations as they are int erpreted by the ethnological theory of
Kulturkreis . A system of mystical correspondence between nature and
man, f->ern·een the elements, astrological signs, the seasons, and so unds,
causes the spider dance, the sword dance, and the music specific to them,
to function (by complementing one another ) as a form of therapy acting
both o n the level of accidental reality (char of the illness produced by the
spider's bite ) and on the level of permanent reality (that of the struggle be -
tween life and death, summer and winter, stillness and mo vement, renewal
and decay ). Within the overall configuration of these mystic correspon -
dences, musical sounds, musical instruments, and dance steps occupy a
,,·ell -defined position. It is by ,·irtue of this position, and of the power it
confers upon them, that music and dance ensure the triumph of recovery,
,,·hich should simply be seen as an example of the ,·ictory of life. In this
,·asr theoretical edifice the real and the svmbolic are closely intertwined,
160 C HAPTER THREE

bur a t no point does Schneider seem ro cast doubt upon the ve ry re al as -

pect of the bite and its cure: "case s ob served in detail are coo frequent, and
we cannot ignore them without running the ri sk of ridicule" (1948, 126 ).
Schneider him self seems entirel y prepared ro accept completely the general
conception of the world within which the cure is effected . Ind eed, he
seems ro see it as superior ro a n y other in pro\ ·iding a n account of the ulti -
mate re a lit y o f thin gs . Too b ad if the ·'c at astrophic theories of Descartes"
(ibid . , 128 ) h av e been responsible for releg a ting thi s conception to "the
camp of poetr y" where "any madnes s is permitted. " Ir is the only one that
contains the truth concerning "the impenetrable myster y of creation"
(ibid. , 126 ). Schneider 's case th us raise s the ver y problem of a particular
mystique a nd its value as a current system fo r interpreting the world. I
shall of course refr a in from entering inro the deb at e. Suffice it ro say char
in terms of our interest here , the symbolic svsrem he de scribes seems to be
the recon struction of a n edifice who se exi stence as a coherent whole has
ne ver acruall v been ob ser ved. The element s th at m a ke it up do indeed a ll
h av e a re a lit y, but the y ha\ ·e been gath ered fro m th e four corner s of the
earth, a nd nothin g pro ves the y h av e e\·er co nstituted a whole.
In hi s L a te rra d ef rim orso ( 196 1 ),' ' \\·hi ch a pp ea red thirteen yea rs after
Schneider 's \, ·ork , Erne sto De M a nin o p rov ide s a rorall v different inter-
pr eta ti o n o f ra ra nrism . H is w ork deri ves to a ver v large extent from the
ream field, ..vork he conduct ed in Sa lem o in 19 5 9 a nd it is in light ofrhe m a-
ter ia ls ga ther ed durin g thi s im ·esti ga ti o n tha t he reint e rprets the old text s
u sed b y M a riu s Schn eider , to \\·h os e rh eo n · h e refer s ver r little (1 96 6,
301 -2 ). D e Ma rti no sees tara nti sm as a "m ino r relig io u s form' ' centered
a round a kind o f "c ho reog raph o -chr o m arico -m us ica l exorci sm " (ibid. ,
7 1 ) a nd as an in stiruti o n th at '·is nor red u cib le to a co mm o n type " (ibid.,
3 0 5 ). W h e rea s he in fac t ac kn o w ledges that its or igins mu st be sought in
th e "o rgi as tic a n d initi aton· cult s of classic a l anr iq uit \'," a nd w herea s h e
co n side r s th e re a re "Afri can a n a logues " (ibid. , 204 ) th at ju srifv a "co m-
m o n M edit err a ne a n cu ltur al ho mel and " fo r the ·::.Z ir, th e bori, a nd the Ibe-
ri a n a nd Sa rdini a n fo rm s of ra r ant ism (ib id ., 30 5 ), he n everthele ss think s
it m o re imp o rt a nt ro str ess it s '·cultur al a u to n Om \·.'' According to him , t a r-
anr ism sh o uld be seen as "a n int ern a l epi so d e \,·ithin an expanding Chris -
ti a n ci\·ili za ri o n " b o rn in the Middle :\g es a n d rela ted to rhe Cru sade s,
durin g \,·hich the "re a l a tt ack s of sp id er p o iso nin g experienced by the
Chri sti an armi es in their camp s" facilita ted "the birth o f the tarantul a svm-
bo l. " H av ing said thi s , h e scares th a t th e essent ia l of raranrism, as an insti -
tuti o n, is to fun cti o n on a ·'mnhico -riru al h o ri zo n of rec apture and reinte-
gr a tion in rel a tio n to critic a l moment s o f hum a n exi stence , with a marked
preferen ce for th e cri sis of p u bert\' , th e th em e o f the forbidden eros, and
the conflicts o f ad o lescence, w ithin th e fr ame \\·ork o f a peasant lifest yle"
(ibid., 304 ). 65
Music, Shamanism, Mediumship, Exorcism 161
De Martino's rhesis rhus seems ro be concerned wirh esrablish in g rhe
hisrorical perspecri ve of rhe dara and deeply impregn a red wirh psychoa n a-
lysis. There would be norhing ro criricize in chis ourlook we re ir nor for rhe
face char ir makes u s lose sighr of whar is, afre r all, rhe mosr obv ious aspecr
of raranrism; char is , ident ificarion wirh rhe spider. One particular point in
his argument, sym promaric of rhe resr, shows chis very clearly. One of rhe
dance figures of rhe rarantulees - rhe best known - consists, as we know,
in imiraring rhe sp ider' s mo ve ments: back ro rhe ground, body arche d ro a
grear o r lesser degree, rhe rarantulee moves about like a spider on all fours.
One can see chis ve ry clearlv in D . Carpitella's film, and the sight is a str ik-
ing one. In certain cases rhe raranrulee utilizes, or rather used ro utilize,
since the pracrice has now died om, another dance figure . When rhe se-
ance rook place in rhe open ai r, De Manino cells us, quoring Kircher's sev -
enteenrh -cenrury accounr, "some tarantulees !er rhemselves hang from
trees by ropes, showing that chis suspension was mosr pleasant for
chem ... ; such a passion affecred above all chose \\·ho had been birren by
rarantulas char had rhe habir of hanging rheir webs from trees" (ibid.,
138 ). When rhe dance rook place indoors, rhe y hung from a rope fixed ro
rhe ceiling. There is every reason ro see in chis beha\·ior a pure and simple
imita ti on of rhe tarantula. Wirhout denying rhe realirY of this imirarion,
De Martino prefers ro see somerhing else in it. According ro him, this fig-
ure "highlights rhe particular symbolic value of rhe swing" (ibid. ). Asso-
ciating chis aio resi s (ibid., 230ff. )- the ace of "!erring oneself hang in
pace'' - wirh rhe mnh of rhe S\,·ing of rhe hanged virgins and wirh char of
Phaedra 's hanging, he interprets rhe hanging of the taranrulees as a sign of
frustrated, unhapp\·, or thwarted female passions. For him a number of
other ymbols can be added ro rhis one : ·'being rocked in a mother's arms";
a game in space char is "always rhe same, at first granted, rhen immediately
taken away"; and finalh·, a "prefiguration of rhe embrace·· wirh a possible
husband. The problem \,·irh such an amalgam is that ir not only proYes
norhing (in face, ir makes one think chat anything can be srmbol ic of an y-
thing ), bur ir also obscures rhe essential core of the marcer, \,·hich is rhar
he re we have people in trance busy identifying themselves, through dance,
with a crearure rhoughr ro have made rhem ill; in other words, char we are
dealing \,·irh a cha r acrerisric case of possession. De Manino i certainly
cor rect in seeing behind this behavior impul ses srrongly ringed wi rh an
ero ticism whose demons need robe exorcized . Bur one could say jusr rhe
same about the Hausa bori, for example . For us, rhe essenrial rhing ro re -
member is nor \,·hat exacrly chose involved are cryin g ro rid themselves of,
bur rather ho\,· rhey do ir.
Seen in light of direcrly observable behavior manifesred by rhe afflicred
person rarher th an from rhe srandpoint of its u·nderl ying representations,
be they real o r conjectural, tarantism displ ays a ll rhe characteristics of
162 CHAPTER THREE

possession as it operates in che gro up of cults recogn ized as possession

cults. In face it dovetails perfeccly well wi th all the points of the definition
proposed earlier:
1 . It cake s the form of a religious manifestation, since it is placed under
che sign of Saint Paul, whose chapel ser ves as a "theater" for che tarantu-
lees ' public meetings. And even though the spide r itself is nor (is no long -
er? ) a div in ity in the ritual, it nevertheless seems co be consta ncly inter -
changeable w ith Saine Paul. The female carantulees (ibid., 363 - 64 ), w ho
still dress as "brides of Saine Paul" (ibid . , 240 ) even toda y sing:

Say where the tarantula stung you

Underneath the hem of my skirt
.. . Oh my Saint Paul of the tarantulas
Who stings all the girls
And makes them saints

2. It is cha r acterized by a change in the behavior of the person con -

cerned, and chis change is accompanied by trance . When she is en-
tranced-and her behavior indicates on every count that she is- the ta ran -
tulee indulges in all kinds of extravagant behavior, of wh ich the most
notorious is that of behaving like a tarantula (that is, of identif ying herself
wich it, as thou gh she were possessed by it ), since her dance imi tates the
spider' movements.
3. Healing is the manifest fun cti on of carantism. But what is it a cure
for? This, as vve shall see, is whe re its very particular nature comes in.
Ultimately, then, what is missing from tarantism co make it a possession
cuJr? Simply daring to call ir one . Why) Because if it had been, it would
have long since ended in the same way as the events at Loudun ended: at
the stake. H. E . Sigerist ( 1948 , 114 ) saw this perfectly well. Whether the
origins of tarantism must entirely be sought, as he thought, in the orgiastic
cults of Greek antiquity, or \\·hether they became mixed w ith African con -
tributions, as De Martino seems ready to admit, is of little consequence.
The essential point is that the Church of Rome could never for a moment
have tole raced its existence as an O\·ert possession cult. The bite of che tar -
antula, whose effects coincide so extraordinarily closely with the signs
that herald the onse t of possession, provided a providential al ibi . As Siger -
isc points ouc, \\·omen \\·ho gave th emselves up to these practices were no
longer sinners buc unfortunate victims of the tarantula. Fundamentally,
then, carantism is nothing other chan a possession cult that dares not
speak its name . This does nor mean ic has no particular traits of its own.
Indeed, it has some very marked characteristics chat der ive, in part at
least, from the fact that it had to remain unavO\\·ed.
Without wanting co dispute the interest of Schneider's and De Martino's
theses, I have tried to identify the reasons that justify integrating raran -
Music, Sham anism, Mediumship, Exorci sm 163
ti sm , without reser vations, into the category of possession cult s. Thi s
task, I hope, is now accomplished. We ca n now mo ve on to exo rcism an d
m USJC.
As we saw, ta rantism is regarded by De M artino as "choreog rapho -
chromatico -musical exorcism" that prim ar ily functions, let us say for sim-
plicity's sake, as a ps ych ic rele ase mech an ism , so char the spide r 's bite,
w hich is more often imagin a ry th an real, operate on the sym bolic level
and provides the opportunity, as it were, for rhe cure . De Manino cites
(1966, 256 ) a fourteenth -centur y text , which is "the most ancient docu -
ment on the musical exorcism of those po isoned by a tar a ntula bite, " that
interprets t aranti sm as a cure operating not on the symbolic level th is
time , but on the level of realit y, since the actual bite is never doubted at
any point. Music and dance are seen as a sort of mechanic al expulsion, as
ir we re, of the venom. In other wo rd s, this text, rhe Sertum papale de ven -
enis, view s tarantism as an exa mple of what was to be known three cen-
turies later, in Kircher 's tim e, as "iacromusic" or curing music, or what we
would term musicotherapy. The text claims th at the ·'joy" the tarantulees
feel at the sound of runes th at please them is ap par entl y re spons ible fo r
their recovery, bur the author adds th at, in realit y, rhe joy terns from the
fac e chat "by me ans of the melodie s and songs their humour s [bod y fluids ]
are dr awn from the interior o f their body toward rhe exterior, preventing
the venom ... from penetrating inw ard s; consequentl y the most impo r-
t ant pans [of their organism) do not suffer from it , but on the contrary are
relieved'' (ibid., 25 7 ). Thi s text is also interesting in that it gives an ac -
count of the popular interpr eta tio n of tarantism (only to refute it ): "the
common pe ople and th e ign o rant affirm chat w hen the t arantula bites, it
emirs music, a nd that \\·hen the sick man hears melod ies or so ng s con -
forming to this music , he then deri\·es grea t re lief from them . Bue in m y
op ini o n rhe situ ation is quire different" (ibid., 256 ). Of the rwo interpreta -
tions chis rexr provides, th e popular and che learned one proposed in its
place, the first is clearly nearer to rhe truth: it alone cakes into ac co unt
\\·hat is, as \\·e have seen, an essential feature of carantism : identification
\\·irh t he spider. Fo r iri s because the melody "confor m s"to the music emit -
red b\· the spider th at ir has rhe virtue of relieving the sick person. We in
face knO\\. char ther e are various kinds of tarantulas, and that the carantu-
lee is o nl y su pp osed to re ac t to the sou nd of th e tarantella specific to the
tarantula that has bitten her. Ind eed, this is hov, th e latter is identified, by
means of the "right rune·· or, if one prefers, the right musical mono, \Vhich
a we saw earlier pl ays such an imp ortant role in oche r possession cults.
This hav ing been sai d , lee us note th at the Sertum does not actually
mention exorcism, either in its own interpretation of t arantism or in the
popular interpret ation it rec o rd s . De Manino is rhe one \\·ho is interpret -
in g whe n he writes th at the Sert11m is a "doc um ent rel at ing to musical ex -
164 CHAPTER THREE

orcism": neither the word nor the idea actually appears in the text. In us -
ing rhe "·ord, De Manino is apparently making himself hei r to a C hri sti an
tradition whose makeup he himself desc ri bes : "The use of ca n o n ic exor -
cism and the elabo ration of cr isis as a means of doing penance w er e rh e ex -
pedients co which the [medieval ] Church sought recourse in o rd er co dea l
\\·irh rhe situation'' (ibid . , 265 ). And it is significant char on t h e same page
of La terra del rimorso we read : "In short, the tarantula appea rs as a 'pos -
sessing spirit' char exorcism controls, ·' and char a few lines la cer we come
across a quotation from a pr iest w riting in rhe seventeenth cen tury who
speaks of rhe Apul ian ra r ant u lees as peop le "possessed by t h e dem o n " o n
whom "exorcism should be practiced" (ibid . , 192 ). Moreover, D e Ma r-
tino is misusing language: from the standpoint of the Sertum, ex p uls io n
certainly is involved, bur it is venom and nor a spirit rhar is be ing ex p elled,
and moreover no r through words bu r rather through the m echa n ical ac -
tion of rhe m u sic and dance. T his wou ld be of little impo rt ance we re it nor
char the use of the word exorcism masks the most important as p ect of the
possession : in raramism as elsewhere, it is nor a conflictual relation with
the deirv char is invoh·ed, bur the partaking of an alliance . The fact char,
depending on the case, chis rel atio n can va ry from alliance proper to a sim -
ple pact of peaceful coexistence, as I said earlier, is seconda ry.
Despite appearances, rhe di vinity responsible fo r rhe possess io n is nor
rhe one cha r is exorcised. On the contrarv, it is the divinit y conce rn ed who,
b\· allo\,·ing the possessed person co identify himself \,·irh him o r he r , pro -
vide s the me a ns of exorcising the illness - real or imagined - from which
rhe person i suffering. In possession cults in \,·hi ch the "thera peutic di -
mension" - co use Zempleni's expression-is an imp o rt a nt fea ture, the
hYsteriform behaviors displa\·ed b\· the possessed person a re nor the disor -
ders fo r which the cure is intended; o n rhe contran·. rhe\· a re a remed\ · for
deeper disorders of a quire different nature afflicting the sick person; char
is to sa,·. for rhe advers it\ · char strike s him. Thi s means char possession
sho uld be understood as a form o f rherapv of adve rsin ·, bringing into plav
an institutionalized hYsteri a or, if one prefers, a soc ializat ion of hysteri a .
.\1u sic and dance are preciseh- the principal means of socia lizing or insriru-
rionalizing chis h\·sreria, b y pro\·iding it \,·irh sre reon·ped forms of trance,
fo rm s char of course depend upon the er of representations constituting
the particular svsrem of a given cult. In rarantism, rhe rarantella (music
a nd dance ) does nor ha \·e the function of curing the raranrulee of her hvs-
reria, bur on the contrary, provides her with a means of behav in g like a
hysteric in public, in accordance with a model recogni zed by all, thereby
freeing her from inne r misfortune . H o\,·? By pro,·iding her wi t h a m eans
of "coming out of herself" and of communicating with the wo rld, with so -
ciety, with he rself. But whether possession constitutes it elf into a k ind of
Music, Shamanism, Mediumship , Exo rcism 165
hyste riotherap y, as in ta rant ism or in the ndbp, o r whet her it rem a in s un -
connected wi t h any therapeut ic functions and di p lays an ext remel y con-
trolled outward aspect, it must in any case be viewed as a respo n se to a
need for communication .
If , in tarantism, the spider's bite has been seen as constituting the ill tha t
sh o uld be exo rcised whereas, in re a lit y, it is the means b y wh ich th is ill is
exo rci sed, this is becau se the ca se is extraordinaril y a mbiguous. A lth o ugh
the bite is most frequently imaginar y, it can, nevertheless, somet imes be
re a l. \Xlhen real, it produces symptoms that are n ot imagina ry at a ll: toxic
re action, pains , difficult y in standing, m u scula r rigidity, somet ime s sex u al
a rous a l, a nd , in a later stage, "a characteristic sensation of burning and
tingling in the so le s of the feet'' (ibid. , 3 17). 66 Following a period of "v ery
m a r ked dep ression accompanied by anx ietv a nd a feel ing of impending
dea th,'' the p a tient then becomes, on the contrary, "agita ted, ver y a n x -
ious, an d sometimes suffering from h a llucin ations. " Some\,·hat like the
datura u se d in the bo rz (]. M o nfou ga -Nicolas ), the spider's \·e nom in duces
the principal s~·mptom s of possession as it m a nife sts itse lf in the system
mentioned earlier. The tingling sensation in the feet and the " intermittent
tre mblin g that occurs princip a llv in the lo w er limb s (so m et ime s displaying
a co nvulsive aspe ct)" ob\·ioush· cons riru re, in add iti o n , an invitation to
dance . In short, rarantism is the \·en· model of th r id eal scena ri o, since
\\·ithin it the svmbo lic and the real, rhe signifier a n d th e signif ied, the
manifest and the latent, rhe anecdote and the hidden m ea nin g , the pretext
and th e profound m o ti\· ati on o r, to refe r to D e M arti no's thesis (ibid . ,
198- 99 ), morso (and re -morso ) and ri111orso, (b ite, reb ite . remorse ), co -
inc id e perfecrh'. \'(l e hould not fo rget, however, that al th ough the purpose
is sometimes to chase our th e venom and ah,·ays to expel \, ·hat it sy mbol -
izes. the intention is not to chase our the spider, bur. on the contrary, to
identih · onese lf to it . \\·hich is do ne a great deal b\· imitating it in a va riet y
of d iffer ent \\-a\·s.
Ther e would thus be no reason \,·hatsoever fo r viewing t a rantism as
"mus ical exorcism ·· \,·ere it not for the fac t th a t a m o ng th e texts collected
b1· De Martino (ibid., 154-56 ), so me can in fact be rega rded as exor -
c ism s :
She has stung YOU
She has bitten 1·ou

:\nd \\'e are chas ing her a11·a1·

. far a11·a1·

a nd
There 11·here she srung 1·ou
She can be killed
166 CHAPTER THREE

Others accomp a ny dance m ove ments expressing the sp id er fleeing or be -

ing crushed . Still others, howe ver, rend to glo ri fy the ta rantula:
Dance, Maria
.-\nd dance hard
For rhe raranrula
Is li\·ing and nor dead
o r else ro minimize its role :
Ir was nor raranrula, nor caranrella
Bue ic was wine from che cask
This is done , h ow eve r, lighth ea rtedl y a nd w ith acknowledgme ncs to Bac-
chus, o r , in other words , with no intention of exorcis ing.
As expressed in the first tw o of these texts, the desire to get rid of the t ar-
a ncula need not necessaril y be interpreted as a \\·ish fo r exor cism. As Jac -
queline Monfouga -Nicolas emphasize s, chere is ah\·ays a certain am biva-
lence in matters of possession , and possession can \·er y we ll be felt as a
burden by the p os se ssed, even though they m ay also derive great sa tisfac-
ti o n from it , or even h ave been cured by it. In Dahomey, a t the end of an
initiati o n , I remember seein g the Sakpatasi (handma iden s of Sakpar a, god
of rhe earth and smallpox ) de lig htedly brushing themselves off . Thi s
meant rhe vodun was lea\·ing th em , I \,·as told. Jean Ro u ch recentl y told
me th a t , for a ce rt a in Songhar adept of his acqua int ance, go ing to a pos -
sess ion sea nce with the prospect of being possessed somet imes did nor give
him the slightest jov . "P laying hor e" fo r a god is not ne cessa ril y an ag ree-
ab le role . E ve n if rhev a re bene\·olent, th e divinities may nor always leav e
as eas ily as o ne mighr like . In many cases the\· ha\ ·e to be coaxed before
they \,·ill agree to go . Bue this does not mean that the\· are being driven our.

To be sure. ·'exorci m" is a \\·o rd char can easih- be used metap ho ric allv.
le is significant, hO\\-e\·er. rhac ir \\·as in an It alian music magaz ine, Mu-
::.ak.6-rhar I read the follo\\·ing statement by a young /\ !oroccan mu sicia n,
the founder of d group that dra\,·s its inspiration from Gna\ \·a music: ·'I
belie\·e in the liberating \·alue of tranc e through music, as a mea n s of exor -
c ising nor the spirits of e\'il bur. s\·mbolicalh-. the spirits of the rn·entierh
centun·. "•·
4 The Strange Mechanism

In both possession and shama ni sm it is clea r that mus ic usua lly pl ays a
role in tr igge ri ng trance . Bur this "stra nge mech anism," as J ea n Ro uch
would call it , has been the object of the mos t di verse, an d so metime s ut-
terly confl ict ing int erpretat io ns. Befo re m oving on to t he secon d pa rt o f
this book, in wh ich we shall look at trance in relation to Greek an ti qu it y,
the Renaissance, and Islam, let us examine the theories through wh ich th e
facts ha ve been viewed thus far . These theo ries gravitate around two
poles: let us call them, using Jean -Jacques Rousseau 's words, that of m u-
sic's moral action and that of music's physical ac tion . Various theori es , we
shall find, combine both aspects in different ways .

Rousseau and the Physical Power of Sounds

Jean-Jacques Rousseau - whom Claude Levi-Strauss (1962 ) has cal led the
founder of the social sciences - could also be seen as the founder of ethno -
musicology, since some of the rem arks in his Dicti onnaire de musique , 1
and the Essai sur !'origine des langues fundamentall y pose for mus ic not
only the problem of the "other" but also of the rel ations between natu re
and culture. It is therefore particularly interesting to note that he himself
wavered between these two poles, th at of moral action and that of phys -
ical act ion (another aspect of the culture / nature opposition ) precisel y
w ith respect to carantism. In Rousseau's rime there was a great deal of dis -
cussion abo ut the curat ive powers of music on the one hand and its phys -
ical action as mechanical vibratio ns on the other. In his Dictionnaire
(176 8, 312 ) Rousseau ,nites: "Although music ha s little power over the
affectio ns of the soul, it is nevertheless capable of acting ph ysically upon
bodies, witness the story of the Ta rantu la, which is too well know n to re-
count here . .. . " Undaunted at having to con tr adict himself qu ite open ly,
he asserts the following in a passage from the Essai, which I feel I m ust
quote in full: 2
As proof of rhe ph,·sical power of sounds , some adduce the healing of
T arantula st ings. Thi s example proves quire rhe opposite . Neither
absolure sounds nor rhe same runes are required ro cure all those srung
bv chis insect. for each needs rhe airs of a melody kno,rn ro him and
168 CHAPTER FOUR

phrases that he underst ands . Th e It a lian must h ave Italian rune s, the
Tur k would need Tur kish runes. E ach ;s onh- affec ted b y accents
fa mili a r to him; one's nerves \\·ill resp ond only to the degree to which
o ne 's mind prepares them for ir: he must understand the language
spoken to him before what he is bein g told set s him in morion. Bernier's
Canta t as a re sa id to have healed a Fre nch musici an's fever; the y would
have caused it in a mus ician of any other nat io n. [1970, 165 ]

E \·en th o ugh Rousseau, like ever yone in hi s da y, rega rded t ara nti sm as a
particularly specracul a r case of he al ing through music , whereas I reg ard it
as a case of possession tr a nce, what he rel ate s in this p assage is th e ve ry
mechanism of the music al mono we de scribed in a pre vious chapter. Bur
for him, this interpretati on of thin gs is onl y one pan of a much vas ter the-
o ry of music as sign. Di scussing sensation s in the chapter "On Melod y" of
his Essa/ (ibid., 14 7 ), Ro us sea u w rite s: "they do no r affec t us solel y as sen -
sat ion s, bur as sign s o r im ages . . . . " Further on, in the chapter entitled
"Th a t Our Most Ac u te Sensations Mai n ly Act through Moral Impre s-
sio ns ," (ibid. , 163 ), he a lso writes: "As long as we choose to consider
sounds o nl y through the commotion th e:· stir in ou r nen-es, w e shall never
have the true principles of music a nd of its power over our hearts. Sounds
in the melody do not act so lely as so und s, but as signs' of our aff ec -
tions . . .. "'
In his Dicti onna ire de musique, Rou ssea u applies this sign the o ry to a
concrete case, that of the famous rune "Ra nz des vac he s, " w hi ch amo ng
the Swiss can trigger \,·ha t I, fo r the purposes of rhi book, h av e suggested
\\·e term emotional trance . Perh aps it is also ,\·orth underscor ing tha t thi s
re ference to the "Ranz des va ches" is p art of a long passage under the entr y
"M u sic," w hich is devoted to what o ne might call, a nticip at ing somew h at,
erhnomus icology, a nd th at the rune is printed at the end of th e vo lume
alo ng \\·irh fou r othe rs, o ne Chinese, one Persi an, and two Am erindian ,
\\·irh the inte ntion, Rou sseau sars, of "enabling rhe Reader to judge for
himself Peoples · va riou s mu sical a ccents." Thi s is w h a t he \\Tires:
I ha\·e added on rhe same Plate rhe celebrated Ranz-des-Vaches, rhar Air
so dear to rhe hearts of rhe S\\·iss rhar pla,·ing ir in their Troops was
fo r bidden on pain of death, because ir made those who heard ir \\·eep .
desert, or die, so a rdent a desire did ir arouse in them to see th ei r
counrry aga in. One \\·ould seek in vain in rhis Air rhe energetic accents
capab le of producing such astonish ing effec ts . Those effects, which
ne\ ·er rake place in foreigners, deri\·e sole!)' from habit. from mem o rie s,
from a th ousa nd circumstances th at, called back b)' thi s Air to those
w h o hear it, and rec all ing thei r co unrn ·, rheir forme r pleasures,
their ro mh, a nd their whole m anne r of li\·ing, exci te in rhem a bitter
so rr ow for hav ing lost all tha t. Th e music is rh en nor ac ting precisel y as
music, bur as a mnemonic sign . To d ay, thi s Air, wh ich is still the
sa me as it ever was, does no r have the sa me effects ir o nce had on the
The tr ange Mechanism 16 9
Swiss; because, having lost the taste for their previous simplicity,
they no longer regret it when reminded of it . So true is it that it is nor in
their physical action that we must seek for the greatest effects of
Sounds upon the human heart . (1768, 314 )

Elsewhere, replying in advance to rhose (and rhev would be numerou s )

who would larer consider the violence of the music. especially the drums,
as the principal stimulus to trance, he clarifies his thoughts even more by
writing : "musicians \\'ho view the pO\\·er of sounds onh · in terms of the ac-
tion of the a ir and the excitation of fibers are far from knowing in what _the
force of this arr trul y resides. The closer the y approach to physical imp res-
sion the further they rake it from its \\·ellspring, and the more also do the y
deprive it of its primitive energy." ' To be sure, Rousseau did nor foresee
the explosion of decibels favored in ou r day by certain types of pop music,
unle;,shed by multi-thousand-watt a mplifiers throu gh loudspeakers that !•• r

have become instruments of acousrico -mu sical aggression .

In adopting this position \\·irh respect to rhe "effects" of music, Rous -
seau \\ ·as re ac ting agains t the ideas of his time-\,·hich he had initially
shared - and forged the way for cultural relativism in the field of music .
("'\"X/hyis our most affecting music no more th an empn· noise to the ear of a
Caribbee 1 " he asked in the Essai. 6 "Are hi s ner ves of a different nature
from ours .... " ) Hi s interpretation of the way music acts in rarantism is
clearly the right one, bur it \\·as hardly listened to, and rhe conviction rhar
in such cases rhe musical sounds act rhrough the '·\·ibrations they excite in
our nervous sysrems'' \\·mild continue to have parrisans. In facr, cerrain re -
cenrlv expressed ideas on rhe effecrs of percussion instruments, particularly
those of the drum, belong to this trend of thou ght. Let us now examine
them.

Percussion and Drums

In an earlie r chapter, ' \,·e considered at some lengrh rhe possibilin· rhar the
triggering of trance ma\· be attributable to a certain su i ge neris pO\,·er
character istic of music al instruments. \"Xiefound, in face. rhar rhis \·ery
\\"ide spread h\·porhesis \\·as unfounded . bur certain conceptions are so
deeph · engrained that I feel it necessan· to return to this poinr once aga in ,
and to look somewh at closely at the case of percussion instruments, in
particula r the drum.
For a \·ariety of reasons the drum is surrounded b\· such a particular
aur a char nor only the man in the st re et bur also, ver v often, the man of sci -
ence readily lends it verv special pO\\·ers . Th e often explosi\·e, viole nt, and
brutal nature of the sounds it produces, a nd the frequenth- dramatic or
obsessive use to which it is put, indispurabh- confer upon rhe drum a par -
ricul a rh· srrong emotional impact. !rs sound can be a truly aggress ive
170 CHAPTER FOUR

force, and its vibrations can have an alm os t palpable imp act. Whether we
observe it in Europe, w here it is an instrument of war, o r whethe r w e envi~
sio n it in tho se distant and barb aric lands in wh ich literature and the cine -
ma inevitabl y assoc iate it wi th the bl oo dy an d tempestuous rites of"p rimi-
tive" religions (pa rticu larl y those cerem onies ofte n indiscr iminatel y
lumped rogethe r under the va gue term "voodoo" ), the drum is regarded as
the instrument par excellence of frenz y. If , ro use Rousseau's term s, there
is one instrument capable of "s haking our ne rves," then it must, o ne
would think, be the drum. Moreover, it is also the instrument p ar excel -
lence of rhythm, and therefore of dance. It is easy ro unde rstand how, car -
ried away by their imagin at ions, some poeple have believ ed th at it is ab le
phys ically, and as it were mechanicall y, ro project people out of them -
selves . H ence the liter ature, often extremelv bad, that has developed
around the drum - the "p ier cing call of the rom -tom " !- and subsequentl y
t he scientific theories, of equally doubtful merit, tha t hav e recentl y been
constructed around it , and wh ich we w ill ha ve to exam ine in greate r detail
later on . Let us begin, however, w ith percuss ion and w ith Rodne y Need -
ham' 8 article "Percuss ion and Tr ansi ti on," which app eared in 1967 in
Man.
Starting from a number of reflec tions on the shaman beating hi s drum
in o rder t0 es tablish contact with the sp ir its, 1eedham ( 1967, 607 ) gr ad u-
ally mo ves on ro the observat ion that "All ove r the world . .. percu ssion
... pe rm its or accompanies commun ication with th e other wo rld." The
problem then, according to h im, is to disco ver what th e exact re latio n is
bern·een the co ncept of sp iritu al ex iste nce and thi s "non -cultural affecti ve
appea l of p ercuss io n" (ibid ., 6 10 ). Why "non-cultur al" ? Because the
me ans used for thi s summons ch ange from one soc iety to another, because
in so me cases the drum is used, in others the gong, and in still others hand-
clapping . Needham, th erefore, srates th at the imp act de rives from percus -
sion itself, and not from rhvthm, melody, or the repetit ion of a p arti cul ar
note or resonance . After observ ing that "t here is no doubt that so und -
\\'a\· es have neural and o rganic effec ts on human beings, irrespective of the
cultural formation of the latter ," he then adds that, of all sou nds , those
th at produce the greatest effect o f this kind are those ob ta ined by percus-
sion, since percussion invoh ·es the "found at ions of aurally ge nerated emo -
tion" (ibid., 611 ). 9 Next, observ in g that drums and percussion instru -
ments are not solely linked \,·ith communication with the other wo rld , but
a re also used in many other circumstances, \,·hich are very often transition
rites, Needham posits "a constant and immediately recognizable associ -
atio n between the type of sou nd and the tYpe of rite," as a consequence of
\\·hich, he argue , "there is a signif icant connection between percussion
and transition' ' (ibid. ). I hall not d\,·ell on the validi ty of thi proposition
here, or on the argument. eedham gi,·es for th e use of fireworks in Chi -
The Strange Mechanism 171
nese marriage ceremonies (explosions that represent percussion at its pur-
est, marriage which is the very archety pe of the transition rice, fireworks
ent irely devoid of any rh ythm ic character ), or on his epistemological re -
flections suggested by the connection he posits. I would like onl y to ob-
serve that, as John Blacking stres sed in hi s answe r to Needha m, 10 the no-
tion of percussion constitutes an ext rem ely d isputable catego ry in
musicology. This alone is sufficient to cast gr av e doubt on the va lidit y of
the connection Need h am has proposed . We know, moreover, th at Claude
Levi-Strauss devo te d a whole section of his work Du miel aux cendres to
the "Instrume nt s of Darkness." In a lette r published by Man in answ er to
Blacking, eedham expresses astonish ment at not ha ving fou nd any con-
firm ation of his views in the Levi-Strauss work . T he re ason is, th at al-
though Levi -Strauss does indeed discuss these instrument s in relation to
"transition", he is careful not to regard the notion of "percussion" - w hich
is precisely what Blacking rejects - as a pertinent one .
Let u s now come to the p oi nt of int eres t to us here. Referring to the trig-
gering of trance in H aitian voodoo, 1 -eedh am claims that it results from
disturba nces brought about by the sounds of the drums "in the inner ea r,
an o rgan which modulates postural atti tu des, muscul ar tonus, brea thin g
rh ythms , heartbeat, blood pressure, feelings of nausea, and certain eye re -
flexes ." '' For eedham, percussion chus should be viewed from a purely
aco ustic viewpoint, and its effects from a purel y ph ysiological one . For
heuristic reasons (which are not worth repeating here and w hich I do not
find very con vincing, moreo ver ), Needham ne vertheless regards the w hy
o f this rel ation between "percu ssion and tr ansition ., as remaining open .
Tw o notes that appea re d in M an the following year, in response to
N eedh a m's article, suggested \\·ay s of extricating him fro m his diff icultie s
by t aking the mechanistic aspecr of his theo ry a st age further. True, both
writers contested the va lidity of the connection made bem ·een "percussion
and tr ansition ," judging it to o gener al to be signific ant, and they \\·ere of
the opinion chat stripping it o f all reference to rh ythm, as Needh am did,
made it \·aluele ss . Fo r it is the rhythm th at is imp ort ant , the y both
cla imed , nor ju st percussion it self. On e of th ese w riter s, ho we \·er , again
reso rt s to eedham 's arguments in o rd er to expla in the effect o f drums on
th e tr iggering o f tr a nce . If the se effect s ar e so \\·idespr ea d. W . C. Sturte -
va nt (1968, 134 ) writes in his letter to the ediror he aded ·'Ca tegories , Per-
cu ssion and Ph ysiolog y," it is because "so me uni versal psycholog ical o r
physio lo gical mechani sm is at work .'' A. Jackson , for his part (1968 ,
297 ), in his article "Sound and Ritual' ' say s th at "sin ce the brain is a com-
mon denominaror tO all mankind, it follO\\' S that wh at is true at the neuro -
ph ysiological level mu st be universall y tru e. " Here we are more firmly en -
trenched than ever in the physical -effect -of -sou nds camp. And the reason
these m·o authors feel so secure in their beliefs is that both are basing
172 CHAPTER FOUR

themsel ves upon a neuroph ysiological theor y of the effects of th e drum

that the y ac cept as pro ven, a theor y according to wh ich the r hythm ic bea t -
ing of th is in st rumenr is capab le of producing a p art icular effect u pon the ·
central nervous system and thereby triggering convulsions.
Shei la S. Wa lker (19 72, 17-24 ), in her chapre r on the neurophysiologi -
cal aspects of possession, lends cenrral importance ro rhi s theory, the va -
lid ity of which she does nor question fo r one momenr. Although in this
same cha pt er (ibid. , 24 ) she prudently confesses she cannot tell "why some
societies have drum -induced possessions and oche rs do not," in her con -
clusion she loses all sight of chat restriction, a nd those fa mous mechanical
effects of drumming a re a llotted , if not full re spo nsibil ity, at least the m a-
jor sha re of it in the triggering of t ra nce : "T he mosr fundamenral elemenr
of possession, " she says, "is the presence of neurophysiological changes,
[a nd the se) are m os t frequently produced b y a sensory bombardmenr, usu -
ally in the fom1 of the son ic driv ing ' 2 of the drum rh yth ms." A little lacer, ob -
se rv ing chat it is sometimes difficult to separate purely neurological ele-
ments from ocher k ind s, she a dd s : " H owever, some time s people seem ro
be reac tin g just to the effects of che drums " (ibid . , 14 8). And w ith refer -
ence ro hypnosis and the importanr role it pl ays in possession, she notes
chat "the hypnotic state is trigger ed by the alte red state of consciousness
and cha nges in bod y ego prod uced by the neurop h ysiological effec ts of the
rh ythmic drumming." It muse be remembered chat Walke r in face as cr ibes
very great importance ro the cul tur al decerminanrs of possession in general.
We see chat for her o n e muse neve rth eless turn ro a neurophysiological the -
ory of the effects of d rumming - th e same theor y invoked by Sturce va nr
and J ackson - in orde r ro find an explanation fo r the triggering of tr a nce.

A Neurop hys iologic a l Th eory of the Effects of Drumm ing

Lee me begin , without preamble, by say ing righr away chat thi s theor y,
\,·hich is advanced b\· Andre\,. Neher, does not stand up to examina ti o n ,
and that it is not necessary to be a specialis t in the field to see \,·h y. How-
ever, che scienr ific, or r ather pseudoscienrific, apparatus with which it is
surrounded has earned it a cerrain credit, as is arres ted not only by the a t-
tention paid to it b\' the rhree aurho rs al re ad\· menrioned, but also by the
pages Ra ymond Prince ( 1968 , 133-35 ) dernced co ir in his a rti cle on en-
cep h alogra ph y and research into possession scares, by the allusions made
ro it during the 1968 Paris Colloquium b\· va rious erhnologis ts, a nd by rhe
reference T. F. Johnston (1972, 30 ) makes co it in his articl e on p ossess ion
music a mong the Tsonga, in w hich it is clear that he, coo, rega rd s it as a
given . ' 3 I therefore have no choice but ro examine this the o ry in fa irl y
grea t detail.
Neher's theor y, the subject of an article published in 1962, a nd the on ly
one ro which the authors just mentioned refer, is based upon th e re sult s of
The Strange Mechanism 173
a number of laborator y experiments reported a year earl ier in a n article
dealing solely w ith neurophysiolog y and conta ining no allus ion w ha tso-
ever to possession. T he very title of the main article ( tha t of 1962 ), "Ph ys-
iological Explanat ion of Unusual Behavior in Ceremonies Involv in g
Drums, " demonstrates the confusion char reigns from the outset. In whose
eyes i< the behav ior in question "unusual") or once is this question even
as ked. The article deals, eher tells us, with this behavior, "often de-
scribed as a trance stare, in which the individual exper iences unusual pe r-
ceptions or hallucinations . In the extreme cases contraction of the body
and generalized convulsion a re reported " (1962, 151 ). The echnograph i-
cal de sc ripti o ns quoted refer to both possession and shamanism.
Th e rhesi is that intermittent aco ustic stimuli having particular charac-
teristics a re capable of "driving" the brain 's al ph a rhythms and there by
triggering con vu lsions. The se particular characteristics a re : ( 1 ) with re -
spe ct to intermittency, a bass frequency corresponding more or less to the
frequency band of the alpha rhythm, which can va ry from eight to thirteen
cycles per secon d , according to the individual; and (2 ) with respect to the
aco ustical spec trum, a predominance of bass frequencies (musical fre-
quencies in this case ) since the y a re capable of tran smitt ing more energ y to
the brain th an higher frequencies without doing damage to the ear.
Neher derived the hypothesis from which this the o ry sprang from re-
search done by various authors on the effects of intermittent light st imul i,
\\·h ich shO\\·ed th a t the latter are capable of causing behavioral distur -
bances and e\ en epileptic fits (Neher 1961, 449; 1962, 153-5 4 ).
Hi s arguments are based on : (1) che re sults of hi s labo ratory experi -
ments using a drum; (2 ) the results of prior experiments done by va rious
scholars on the effects of light stimuli; (3 ) musical rranscriprions o f Afri -
can drumming a nd recordings of Haiti a n music; ( 4 ) erhnographical de -
scriptions . This rag b ag of arguments, put forn·ard in a very confused way,
hardl v inspires confidence, but here agai n, lee us move on .
For his laboratory experiment, Neher (1961, 449 ) used a drum delive r-
ing a signal of lO\\' frequency (75 - 1-0 kh ) and high volu me (120 db ),
struck at the relativel y low frequency of three, four, six, and eight beats a
second, because "it was difficult to beat the drum faster"! The two pub -
lished encephalograms (ibid., 451 ), each of a different brain, were ob -
tained by using a beat of four strokes per second , as the trace indicates.
The y sho\,-, eher says, that the result s of the stimulation were, in both
cases, (1 ) a "driving" phenomenon; and (2 ) blinking of the left eyelid.
The blinking was observed in half of the ten subjects caking part in the
experiment, the driving phenomenon having been observed in all of them.
It seems, however , tha t the results ob taine d with the faster speeds of six o r
eight strokes per second were not good, sinc e the loudness of t he signa l
decreased at these frequencies, "which were not easy to attain" (ib id.,
449 ).
174 CHA PTER FOUR

On the stren gth of these exper iment s, Neher feels he is justified in con -
cluding th at : (1 ) the responses he obtained in his laboratory with audito r y
sti mul ati on are "simil ar " (ibid., 451 ) to those obtained by others with
"p hotic driving"; and (2 ) they confirm the proposed; hypothesis, namely,
that the "un u sual behavior [read: the co nvulsions] obse rved in drum cere-
monies is mainl y the result of rh ythmic drumming wh ich affec ts th e cen -
tral nervous system" (1962, 159 ).
I should like to point out first th at whereas, according to eher (ibid .,
154 ), light stimulation ha s indeed brought on behavioral disorders th at
can reach the stage of "cl inic al psychopath ic states and ep ilepti c seizures,"
he, Neher, has managed to produce noth in g more than "involu nt ary eye -
blinks" ( 1961, 450 ). E ven invok ing, as he does, Fischer's Exact Probabil -
it y T est, there is some abuse, I would think, in regarding these result s as
being of equal va lue, and in claiming , as a consequence, that in o ther con-
d iti ons (perhaps by hitting the drum quicker' But this is not easy! ) he
would have obtained the same re spo nse s, or in any case responses th at
could be diagnosed as "musicogenic epilepsy" (1962, 153 ). Second, 1there
is no reason to accept without further proof that the behavio r al disorders
observed in a laboratory under the effect of intermittent stimuli of any
kind, visual or auditory, are of the same natu re as those obse r ved in p os -
session tr a nces. T here are man y different kinds of convuls ion. Neher
makes reference to epilepsy. In Senegal (Andras Zempleni, personal com -
munica ti on ), the first duty of those responsible for ndop sea nces is to as -
certain, in the case of nonrirualized crisis, whethe r it is the result of epilep -
sy or, on the contrary, at tri butable to possession. Th ird, a priori, the kind
of p arameter s chosen for the stimul ation casts doubt up on the experime n-
tal val ue of the procedu re . The auditor y st imu li used in the laboratory, be -
ing totally constant in form and intensit y, have in practice ver y little in
common w ith the constantly va rying stimuli provided by dru ms pla yed in
possession seances.
All this should suffice, I believe, to arouse extreme skepti cism wi th re -
spect to Neher's theory. But there are even more decisive re aso ns for being
skeptical. Neher base~ his arguments, as I ha\·e said, partly on his own lab -
oratory expe riments , partly on ethn omu icological data . I shall st ate later
\Yhat I think of the latter. Be that as it may, hO\\-e\·er, Neher considers th at
(1 ) a priori the frequency most likel y to indu ce the driving of the alpha
rh yt hm lies in "the range of slightly below 8 to 13 cycles a second" (1962,
154 ); (2 ) in reality, as described in ethnomusicological documents (musi -
cal transcriptions and sound record ings ), '·agitated behavior occurs \\·ith
drum frequencies that re ach o r 9 cycles per second' ' (ibid .); (3 ) he has
obtained results himself with d ru mbeats of four cycles per second. Thi s
means , ultimately, th at the phenomena in question can occ ur when drum s
are struck at a speed varying from, let us sav, twel ve beats per sec ond to
The Strange Mechanism 175
four bears per second; in other words, when they are played at anyw here
between MM 240 and 720. These cadences cover the whole spect rum of
tempi from moderato to prestissimo and be yond . So, unless it is slow,
drumming of any kind must therefore be able to trigger "driving ." In othe r
\.Vords, ever y time a drum is played, or almost, we should expec t to see
people go into convulsions. At least this would be the normal conclusion
of 1eher's arguments. Needless to say, this makes no sense . If eher were
right, half of Africa would be in a trance from the beginning of the yea r to
the end.
Let me add, just in case it seems necessary, that ( 1) the musical tran -
scr ipti ons he in vokes to support his the ory relate to ceremonies that have
nothing to do w ith possession and in wh ich, natur ally enough, no trances
or convulsions ever occur; (2 ) although the recording s he cites (made by
Courlander, which is no coincidence, as we shall see ) are indeed of posses -
sion music, I could cite dozens of others, all with the same rhythms, that
are not. As for the ethnographical rexrs, there is no point in even talking
abou t them : not one of chem pro ves an ything at all about rhe subject un-
der discuss ion, and some are quite simply ludicrous. But the quotation he
takes from Courlander (1944, 45 ) is nevertheless significant. On rhe sub -
ject of drums and possession in Haiti, Courlander indeed says 1 ' chat as a
result of preconditioning, "whenever these rh ythms are heard possession
is induced." Clearly, texts of this kind, and there are a great man y, inspired
Neher's hypothesis. I have no intention of criticizing Courlande r, whose
\York is justly a nd unanimously respected. I only wish to emphasize that
such information must be used with extreme prudence a nd chat thi s is pre-
cisely \\·hat Neher fails to do.
I have dealt with chis theor y at length, I wo uld like to repe at , only be -
cause ic h a s sometimes, quite wrongly, been taken seriously. Bue because
_ 1eher's theor y is va lueless does nor me an chat we are not ju stified in ask-
ing ourselves whether the so und of drums, which does incontestably have,
at least in certain cases, a genuine physical impact upon the listener, is not
capable of inducing certain perturbations of the physiological nervous
state and thereby contribute to trigger ing tr ance. Others besides Neher
have thought so . Dr . Charles Pidoux, for instance (1955, 278 ), a ph ys i-
cian who has been acquainted with possession cults in Mali for a long time
and who is also an ethnopsychiatrist, hypothesized th at drumbeats might
act upon "d ifferent levels of the neural ax is." The research wo rk o n w hich
he bases his ideas has unfortunately nor been published, so it is impossible
to make any judgment on the matter.
It is my view that given the present state of our knowledge, contrary to
w hat may have been said, there is no va lid theor y to justify the idea that
the triggering of trance can be attributed to the neur ophysiolog ical effects
of drum sounds. This does not me an that drumming is never responsible
176 CHAPTER FOUR

for entry into trance, it means only that when it is res p on sible it is so fo r
reasons of anothe r k ind.

From H ypnosis ro Cond itioned Reflex and from Emotions to D rugs

Let us now lea ve theories concei ved in the abstract and pass on to th ose re -
sulting from the observation of trance in the field , and even, som etim es,
from personal experience of it. How do those who have observed and de -
scribed trance concei ve of the relations that ma y exist be tween music an d
trance on the ps ychoph ysiolog ical level? Wha t explanat ion do t hey offe r,
at this level, fo r that "stra n ge mechanism" whic h is th e tri gge ri ng of
trance~
Some of them regard music as ha ving hypnotic action upon the ner vo us
system . For instance, Nin a Rod rigues, a Brazili a n docror of m ed icin e
strongly influenced by the wo rk of Charcor a nd J a net , and w hose wr iti ngs
a re a milestone in the swdy of the candomble w rote eig ht y yea rs ago on
th e "so mnambulism brou ght on b y the hallucinat ion of sac red m usic"
(Rodrigues 1935, 122 ) a nd on "the unu sua l monoton y" of the d ru m ming,
"capa ble of riv aling all the va rious mean s of inducing hypnos is by fat igue
and concentration" (ibid., 111 ). Moving from Braz il to South Africa,
H. A. Junod reg a rded entry int o trance as the result of hypnotic sugges -
tion exerted by a "noisv co ncert " (19 13 , 44 5 ) ro \\·hich the p atient was
subjected h our a fter hour, a concert in wh ich drums a nd rattle s combined
ro guarantee the continuity o f a terribl y int ense sound, a "frightful hub -
bub," o r "infern a l" o r "h ide ous din ." Although, as we saw, he also w rites:
"Th ar w hich is th e m ost essentially nece ssary is the singing, the human
voice .. . " (ibid., 441 ), a nd a lth oug h he recog ni zes the import ance of rhe
meaning of t he \\·ords, it is clear rh ar for him rhe music act s largel y by
overload in g the nen·ous s,·srem, bv a ph\·sical harrass m ent, one might
say, that redu ces t he p atient to a state of exhaustion and ner vo u s cri sis.
Still in the same line of th oug ht, th e mu ch more recent desc ription that
Kate rin a Kako uri (1 965, 21 , 25 ) gives of th e tr ance of the Anastenar ides
in Thr ac e , also pr esents music as ac ting directly on the ner vo us state of
t hose in volved , w ith the "heavy sound" of the drum an d its "per sistent
bea r .. which gets on the nerves of the initi ates or on th e "s tr a ined ner ves of
the possessed ." Ir is indee d the ph ys ical acti on of sou nd on the ner vo us
sys tem that is openly inv oh·ed here . Bur behind other, much less explicit
fo rmul atio n s, such as rhar of R. Firth (1969, xiii ), w ho refers ro the "mo -
notonous bearing of drum ," o r th a t of D . Carpirella (1966, 360 ), who
speaks of rhe "o b sess i\·e rhnhmic repeti t io n" o f rhe tarantella, one can
st ill percei\·e rhe sa me idea of ln·pn os is as resul tin g from the manipulatio n
through music of rhe nervous sys tem o f the per son who will be or al ready
is entranced. M a n y more sim ilar example s could be cited.
The Strang e Mecha nism 177
T o co n sid er t he mech anism of entr y int o tr a n ce as a condit ione d refl ex
t h at is d ue to a mu sical stim ulus is a qui te d iffe rent id ea . We ow e th is the -
o ry to Her skov its, w h o de vel ope d it w ith in the fra m ewo r k of his ge n er a l
inte rpret ati o n o f p oss ess io n as it can be o b serve d in th e candomble o f Br a -
zil. R eac tin g agai n st th e views of N in a R o drigue s and her su ccesso rs , w ho
tr ea t ed trance as a psycho p atholog ical ph en o m enon, H erskovits tho u gh t
of reinse rtin g it in to t he ge ner al sys t em of religio u s rep resenta tio n s char a c-
te ri stic of that socie ty. He wanted co show in chis way chat possess io n
tra n ce sh o u ld be viewed as a no rmal scare res ultin g from app renti ces h ip to
a cultural model, itself la rgely dete r mined by h isto ry.

The psychologic al process we have in vie\\. is rhar which is very clearly

defined by rhe expression 'co nditioned reDex,'" \,·hich means rhar
ewn· rime a specific stimulus is applied, there is a corresponding
reaction, rhe individual ha,·ing been accustomed ro beha,·e in this ,,·ay in
respo nse ro an agreed-upon signal. Ther e is nothing ab normal in this
process - qu ite the contrary, since it represents the p Ychological level
at which a large pan of our existence is lived . ow let us imagine a
person who has been brought up in a cultural environment in ,,·hich
there is a profound belief in divinities, and in ,,·hich he has been taught
since childhood that he will receive, or be capable of recei,·ing, one
of these di,·inities; th at these deities are summoned b,· the intermediar v
of spec ific drum rh ythm s and chants, ro which they respond by
descending upon the heads of those chosen ro sen-e them. There is a
good ch ance that, in the presence of rhe stimulus constituted b)' all the
facto rs of a gi\·en situation conforming to rhe indic ations I ha,·e just
gi,·en, the response \\·ill nor be long delayed, and rhar possession \,·ill
rake place. [Hersko,·irz 1943a. 25 ]

The broad lines of this theory \\·ere taken up by Roger Bascide (1945 )
t\,·o Years later; bur Basride places even more emphasis than Herkovirs on
rhe importance of the situation . ''The music does not inevitably lead to
trance," he wri tes (1945, 88 ), 16 for "the selfsame rhythm th at on a day of
ceremony will produce a pos sess ion crisis in an indi\·idual, will produce
nothing if he hears it outside th is context, since the stimulus of the condi -
tioned reflex, as described by Hersko\·irs, is not a physical stimulus, the
hearing of a particular rhvrhm, bur a psychic stimulus, the rh) ·thm assoc i-
ated ,,·irh a certa in day and \\·i rh a certa in place. Ir is also necessary fo r th e
initiate to have had his body purif ied bv certain herbal baths . A set of fac -
tors, regulated by society, must all be present together, o t herwise the m u -
sic w ill have no effect at all."'· This being so, one would expect Basride to
rejec t the notion of conditioned reflex in this context. This is not the case,
ho,,·e, ·er. Despite the re rricrive clause chat he inserts into its opera ti on
an d ch a r he opposes to H ersko, ·irs 's thesis, when he writes ( 1955, 501 ) " it
is not a stimulus [music ] that determines trance .. . it is the total situatio n
178 CHAPTER FOUR

that acts .[it] is the total situat ion that causes music to lead to tr ance,
or, on the contrary, prevents it from producing crisis," he still remains
faithful to it, for in his last writ in gs on trance (1972, 73 ) he st ates that it is .
because of "condi tioned reflexes" char, once initiated, the adept can go
into trance again "upon hearing certain musical leitmotifs."
We still must ask, nonetheless, if it is in fact legitimate to speak of cond i-
tioned reflex in these circumstances, since either it re ally is one, in wh ich
case it ought to occur in response to its stimulus in any situation, or else
the stimulus sometimes does nor obta in a response, in wh ich case it is not
a reflex of this type that is involved. The question is an import ant one,
since it involves the level of consciousness at which the mechanism of en-
tr y into trance operates . I m yself am inclined to chink rhar if the trigger in g
of trance by music is subject to so m any circumstantial restrictions - and
observation undoubtedl y confirms that ir is- then to talk of conditioned
reflex merel y adds to the confusio n. Henc e the rather oblique use that has
been made of this term. Thus Pierre Verger (1 969, 59 ) tells us th at rh yth-
mic motifs pla yed on drums are the basis of "a sort of conditioned reflex "
acquired during initiation. For Jacqueline Monfouga -Nicolas "the mu -
sico -gestural whole" (19 27 , 189 ) constituted by possession dance should
be rega rded as rh e "condit ion al stimulus " enabling trance to become a
"condit ioned reflex," bur elsewhe re she obse rves rhac "the music does not
induce the trance itself bur rhe form ir must rake" (ibid., 197 ). Margaret
Field (1969, 7 ), for her part, uses rhe word "conditioned" but without as-
soc iating it with refle x . Ha ving noted that in Ghana "dru mming , singing,
clapping, and the rh ythmic beating of gong -gongs and rattles, alone or all
together, are the commone st inducer s of trance," she ad ds: "Most people
who are possessed as p an o f their profession - priests, diviners, priestl y
auxiliaries, and medicine-men's auxiliaries - are conditioned 18 to become
possessed when chey pl ace themsel ves suitably, just as most of us are co n-
ditioned to fall asleep in a co mforrable bed in a dark, quiet room. Some
priests ha ve only to enter the sanctuary when the drums and gongs are
beating and the flute wailing. Some diYiner s gaze int o a black liquid to the
sound of gongs. Another m ay have to hold upon his head, again to the
sou nd of gongs, a \Veight heavy enough to m ake his neck ache ." 19 Bue
there are o cc asion s w hen the spirit, like sleep, is woo ed in va in.
Other writers have, more o r less explicitly, made use of var iou s combi -
nations of the conditioned reflex theor y inherited from Herskovits, th at of
che neuroph ysiological effect of drummin g taken from eher, and that of
the total situation taken from Bascide. Thus Gisele Cossa rd , whose inter -
pretations are especiall y interesting because chey are based on the personal
exper ienc e of being initi ated into the candombl e, on the one hand writes
(1967, 17 5 ) that "a t the rime of initiation a correspondence was estab -
lished between the rh ythm and che trance state," and on the othe r (ibid.,
162 ) th at "o ne is inclined to \\·onder if there is not some ph ysiolog ical ac -
The Strange Mechanism 179
20
tion of certain sounds on the organism. " The reference she then makes to
"low -frequency vibrations" echoes the Neher theor y," although she does
not name him. Indeed , she then goes on to sa y that "whe n the person goes
into trance he makes the characteristic ges ture of putting his hands over
his ears as if to protect himself from the sou nds besieging him ." Neverthe -
less, for Cossard music ultimatel y is only o ne element in a whole set of
stimuli m ade up of "a uditor y sensations: rh ythm of drumbeats , song,
rin ging of the adza bell, cries of greeting to the orlsha; sensatio ns of taste
and smell: smell and tast e of the abo;" vis ual sensations : sight of the con -
secrated objects representing the orlsha o r used to ado rn h im" (ibid.,
192 ).
Ano t her categor y of explanation highlights th e disturbances music -
and also dance - produce in the inner ear . For Zempleni, in the ndop,
whe re "the ritu al cri sis is the normal termination of the d ance and collapse
is the natural conclusion of the 'crisis"' ( 1966, 414 ), the "p hysiological
trigger of co llapse" is "ves tibul ar autos timul ation," w hich , "ow ing to the
increasing volu me and p ace of th e mu sic," leads or rat her "d rive s" the pos-
sessed person ro "extreme muscular exh aus ti on and spa ti al disor ient ation"
(ib id., 417 ). 23 This "spatial disorient ation" thus seems robe caused partly
by music acting (at the physiological level ) on the ear by mea ns of its vo l-
ume and intensit y, which can attain an "a lmost intolerable" level (ibid.,
359 ), and partly by the set of m otor behavior patterns cente red arou nd
dance and cons tituting w hat Ze'mpleni calls "th e labo r of crisis" (ibid .,
415 ).
The interp retation proposed by Fr anci s Huxle y in rel atio n to H ai ti an
voodoo is very simi lar. For him, trance is largel y due to perturba ti ons in
the inner ear produced by drummers w ho m an ipul ate the music until the
"b u ffet s" of the sound have achie ved their full effect (1967, 2 86 ), whi le
drumming, d ancing, and singing combine in orde r to ac hie ve these pertur -
bations th at would set off a process o f the dissoci at ion of conscio usne ss
(ibid., 2 87 ). Howe ver, he also says th at the drummers succeed in settin g
off this dissociation only in those who are "r eady" for it. In othe r wo rds,
Hu xley introduces a restricti ve clause simil ar to th e on e asse rted by Bas-
tide when he wri te s that if a set of conditions is not comp letel y re aliz ed ,
"music has no effect at all." For the same reason that I doubted earlier the
existence of a "conditi oned reflex," I am now skept ical abour the realit y of
these "perturb ati o ns of the inner ea r": either th ey ex ist and do in fact in-
duce th e trance, in which case they should have the same effect in all cir -
cumstances, o r else they do not always h ave thi s effect, in wh ich case why
re sort to them as an explanation, since it is in any event no more than hy-
pothe tica l? We mi ght as well do away w ith it.
I should point out, howeve r, th at th is theory of trance as being caused
by pe rtur ba ti ons of the inner ear was adva nc ed as early as 1948 by Dr.
Aubin , a physician, but only in relation to the dance element, which he de -
180 CHAPTER FOUR

scribed as including "moveme nts of neck and torso both resulti ng in fren -
zied mo vement of the head, which is subjected to truly vertig inous gy r a-
tions and repe ated shaking in every direction. " "This orgy of movements;,
results, he says, in "a violent excitation of the labyrinth." "We know by the
same token that it [rh e excita ti on] eng enders, from the psychic point of
view, a special stare of exaltation,"'' concludes the autho r who, as we can
see, co n found s re aso nin g and tautol ogy .
H owever, I had better m ake my position clear. Th oug h I am skeptical
about perturbations of the inner ear, I do nor neces sarily deny rhar certain
movements (particularly rota ting the head or spinning round and round
on the spot ) do in fact produce them. Wha t I am di sp uting is that these
perturbations should be seen as constituting trance , and that they are held
to be a utom atic. In rhe zar cult, the g11rri(the d ance in rel ation to which
Leiri s quotes rhe p assage from Aubin given abo ve ) marks "the departure
of the sp irit as well as irs arri va l" (19 5 8, 18 ); in o rher wo rds it is associated
as much w ith the end of th e tr ance as with its onset. Thi s fact alone is
proof enough that it does nor inevirab lv prod uce loss of balance in those
\\·ho perform it.
Ir is, moreo ver, necess ary to distinguish between rwo aspects of these
perturb ations of the inner ear. One is linked with the mo vemencs just de -
scribed, and within the limitations I ha ve mentioned there is no reason to
doubt their existence; the other is at tributed nor to the mo vements made
by the subject bur to the sounds he is made to hear. The intensit y of these
sounds is such, the argumenc goes, thar rhe y engender an "ove rsaruration "
of the ear that causes pe ople to lose their bala nce. Thar is the theor y. I will
nor repeat rhe objection I made to ir in an earlier chapter , and which ques-
tions why this oversarurarion affects onlr one p articul ar person and nor all
those aro und him. Thi s objection alone demolishes rhe thesis. W hat I
think ought to be stressed her e is the very subjective nature o f this theor y.
Everyday experience tells us that such incensirv of sound judged to be in-
tolerable by some people can be conside red extremely desi rable by othe rs.
Witness our present -d ay pop concerts, in wh ich rhe sound level, to the de -
light of the fans, reaches a threshold rhar ;s to me pe rsonally (in all my sub -
jecriviry ) intoler ab le. Ir is to be fea re d rhar in judgi ng the "violence " of the
drum min g that so often acco mpanies possession dances, those w ho see it
as a factor in producing sensor ial perturbation are only interpr eting the
facts in a ver y subjecti ve manner.
1\longside , o r in opposition to, the various theories I have listed so far,
all of which claim the onset of trance results, at least in pan, from the
physical ac ti on of music on rhe ner\'OUS S\·srem, there are othe rs that m ake
ab olurelv no use of this kind of explanation. T his holds true particularly
of rhe the.sis of Jean Rouch, for \\·horn the onset of trance results from psy-
chological much more than physiological manipulation of the subjec t by
The Str ange Mech ani sm 181
the zima; that is ro say, by the priest-musician officiati ng a t the ritual.
Th a t R o uch should opt for such a n interp r etation is m a de a ll the more in -
teresting by the fact that possession mu si c among the Songh ay is, in fac t ,
ch aracte rized by ve r y v io lent drumming . ' 5 Thi s drumming is provided, as
we know, by large go urd s upturned on the sand and fitted w ith a device
that incre ases their re sonance . 16 Th e drummers play them wi th sticks a r -
ranged in a fan for m ation, w hich enables them ro o bt a in percussive effects
of unusual inten s it y. As fo r the dance, it ro o is ch a r ac terized b v violence,
a nd ofte n entails rotation of the he ad a t the neck. All the se conditions
might we ll prompt an o bser ver ro evoke perturb a tions of the inner ear
caused both by sensorial oversaturarion and ph ys ic a l overexcirarion. Yer
R o u ch does n o t o nce a llude ro any phenomenon o f thi s kind. H av ing ob-
served that "a pos session dance can las t for sever a l hours, eve n for seve r al
days, w ith o ut success " (w hich excludes, ipso facto, that the dancing and
music mech a nic ally produce a n effect ) , he goes on: "Bur when the zima
know their job well, these failu r es a re rare: the music , the words of the
morroes, the steps of the d a nce, trigger the strange mechanism" ( 1960,
148 ). There then follO\\'S a description of entry int o tr a nce that I must
quote in full:

At an imperceptible sign, the priests and musicians sense that one of

the genii summoned is beginning ro manifest itself (the y feel "a cool breeze"
on their faces ); at th is, the y dispense wirh rhe usual order of the
d ance s, the band now plays o nly one theme, rhe zima recit e only o ne
motto. The dancers as a group cont inue rhe steps the y are performing,
of course, bur o ne of them, an d he alo ne , has now become th e
objec t of the pri ests' solicitude; they surround him and recit e
increas ing ly efficacious mottoes ro him while poinr_ing altern ately ro his
left and right: rhe genius must come neither to rhe left nor to the
right, bur in rhe middle, onto the dancer himself. Abruptly, the latter
qui\·ers, weeps, and halts. Alread y his movement are no lo nger th ose o f
dance bur those of convulsion , and \\·hen rhe trembling reaches
paroxysm the dancer rolls on rhe ground and howls : rhe dance is O\·er.
At this \·ery moment a god incorporates himself in the man's body .
\Xlhen he has been calmed somewhat by rhe attentions of rhe 'qu iet
women ' and risen to his feet again, the H aley's behavior, voice, and
gait become comple te ly different from wha t th ey were before his
possession . The y are char acteri st ic of rhe genius present within him, an d
enable thi s deity ro be recognized with un failing accuracy: for
example, the paralytic Hargey sray on their knees and move about by
hopping on them ; Kyire;\ \,·ho has only o ne eye, keeps one eye
closed; Zatao covers his he ad \,·irh dust and eats earth; :--.·ya/ya,who is a
coquett e, on rhe contrary sh akes off rhe dust char soils her; Sadvara,
who is a snake crawls on rhe ground; Dong o, rhe spiri t of thunde r,
poi nt s at the heavens and groans . .all specific pat tern s of
182 CHAPTER FOUi

behavior: myth materializes in the most singular way, its invisible

characters appear to all the world as they are. And so that this identity
can be even more complete, the genii themselves are the ones who
demand their ritual vestments and props, in that [type of] voice which
also is halting, alien, and distant.
From the \·iewpoint of our ·present subject, what is striking in this de
scription is that Rouch lays no stress either on the frenzy of the danc
(even though he does describe it a short while before ) or on the violence o
the drumming. What he highlights is the fact that the band concentrate
on a single tune and that the zima only recite the one motto. Tune o
mono, it comes to the same thing: a musical message, which is a sign an,
which has a psychological, nor a physiological impact on the subjec
whose entry into trance is being prepared. In fact this descript ion migh
have been made to illustrate our Rousseau quotation earlier: "for eac1
needs the airs of a melody known to him and phrases that he under
stands ." As for trance, it is described as being made up, first, of a crisi,
\,·hich is an emotional paroxysm, then of a change of identity, which, c
course, also has much more to do with psychology than physiology .
Another example of an interpretation of the role of music that makes n,
reference to any kind of "physical power of sounds" is provided by an Al
rican ethnomusicologist, J. H. Kwabena Nketia. Talking of possession i
African society in general, bur w ith particular reference to Ghana, h
writes : "it is belie\·ed that the state of ecstasy or of possession can b
quickly induced and sustained by means of special music closely correla1
ed with specific forms of bodily action . Ir is believed also that the gods ar
sens iti ve to this music. Opportunities a re, therefore, sought to call r,
them while the dancing is going on, in the hope that they will 'po ssess' rh
dancers as they are emotionally prepared to receive chem" ( 19 57, 5:
Here, then, music and dance act in conjunction to produce an emotiom
stare favorable to possession.
Curiously enough, the emotional component of trance plays almost n
part at all in interpretations that have been made of the onse t of posse!
sion, and Nketia is an exception in thinking, as he does, that music pave
the way for it at an emotional level. On the contrary, one of the princip,
accounts we have of shaman ism, th a t of Shikogoroff, places a great deal o
emphasis on this component, and presents music as contributing in a dec i
sive way to this "growth of emotion" that leads to tr ance. Bur here again
as we saw,, - the restrictive clause we have already mentioned twice come
into play: the shaman still has to will his trance .
In the second part of chis book, in the chapter concerning the Arabs, w
shall be dealing at length with musical emotion as a factor in triggerin
trance. For rhe moment I w ill limit myself to unde rscori ng the fact th ,
with this emotion we enter the psychological dimensions of trance, and -
The Suange Mechanism 183
to use Rousseau's terminology once aga in- th at in rhis case it is the "m ora l
action" of the music th at is at wo rk , not its ph ysical action .
Ultimatel y, then, one might search in va in for the re asons wh y the ro le
of music in triggering the "st range mechanism " has so often been viewed
as ph ys iological in nature and, consequently (rhe fin al point I w ish to
m ak e in thi s chapter ), as more o r less comparable in its acti o n tO the use of
a drug. Yet this is implicitl y the int erpretation th at Lew is, for example, ar-
rives at whe n , in listing the va ri o us me ans u sed to indu ce tr ance, he comes
to music (an d dance ) just after ha ving cited "rhe inh alati on of smokes an d
va por s'· and ju st before going on to "the inge stion of such drug s as mesca-
line o r lyserg ic ac id " (1971, 39 ). 1o r is it by chance th at we re ad, in a p as -
sage de voted by Len o r a Greenbaum ro pos session in sub-S aharan Africa,
that possession tr ance "m ay be induced by dru gs, music, or other methods
extern al to the indi vidu al" (197 3, 42 ). T o express thin gs in this way is to
imply that the act ion of music is of the sa me order as that of a dru g, w h ich
does, incontest ab ly, ac t in a "phys ical" way. Alth ough it is pe r fectly per-
mi ssibl e to say, met aph o ric ally, th at music is a drug , in the present con-
text, wh ich is not th at of metaphor, it simply contributes to gene ral confu -
sion .
I hope I have demonstrated by now tha t nothing autho ri zes us to think
that music - at least in sofa r as it is being heard , not made, wh ich is the
case in possession - pl ays any di rect role in the onset of trance other than
by means of its "mo r al ac tion ."
 PARTTWO

Ir is through integration rhar the qualit y of thing s ch anges .

Fr an<;ois J acob, La log ique d u vivant
5 Music and Trance among
the Greeks

In ancient Greece, trance, although this is too often forgotten today, con-
stituted a very important aspect of religious life. Dion ysiac practices, with
their wild beha vior and violence , do of course form an integr al pan of
the picture we tend to have nowada ys of Greek religion. Rohde, and
man y others after him, h ave given accounts of these practices . ietzsche
celebrated them at length. Even though Jeanmaire and Dodds (re spec -
tively in Di onysos and in The Greeks and the Irrati onal , both published in
19 51 ) m ade it quite plain more than thirty years ago, we still too often ig-
nor e the fact that possession trance was the mainspring of the cult of ·
Dionysus -Bacchus . That Bacchus wa s the god of the vine and wine is obvi-
ously no mere chance; to get drunk is ultimately no m o re than a particula r
way of no longer being oneself. But Dion ysiac frenz y can no more be re-
duced to drunkenness than Dionysus can be reduced to Bacchus, contrary
to what the current usage of his latinized name would lead us to believe.
From the standpoint of the histor y of religions, the feeli ng of elatedness
due to w ine is no more than an anecdotal and fairly recent aspect of the
much more ancient, and much more universal elatedness due to trance.
Moreover, religious frenz y was not limited among the Greeks to the wor-
ship of Dion ysus. Ever yo ne knows th at the actual functioning of the Del-
phic oracle, anothe r essential aspect of Greek religi on - this time linked
w ith Apollo, god of music-relied in part at least upon the practice of
mediumistic trance. In addition, like so many other peoples, the Greeks
regarded the inspiration of the poet (an d of the musician, since the two
were for a long while inseparable ) as being nothing other th an a trance. Fi-
nally, Plato, as we shall see, held that there was no true lo ve other than
mad love, and th at this madness, also stemming from the gods, was a
tr ance .
It is of particular inte rest for us to examine the relati ons between music
and trance in ancient Greece more closely than it has been hitherto. First,
because of the importance Greek civilization has for us. Second, because
of its key position, both in space and time: at the juncture of archa ic soci-
eties (those that have constituted the main areas of our observations up till
now ) and the Middle East, hence the Arabs, who undoubtedly owe certain
aspects (recent ones, actually ) of trance as practiced by the Sufis to the
188 CHAPTER FIVE

Greeks, bur who, above all, ha ve profoundly affected trance as it is ob -

served today in large sections of Africa, both black and white. Finally, be-
cause we are indebted to Plato for t he most ancient theory of the relations
between music and trance. A fascinating theor y, moreover, first becaus e of
the stature of its author, bur also because it has the unique quality of being
at one and the same time the theory, first, of an observer who certainly did
nor practice trance himself and only half -believed in possession, but who
belonged to a society in wh ich it was an ever yday occurrence; and second,
of a philosopher , one of whose major concerns was to integrate the irratio-
nal into that essent ially ration al syste m of th e world that he strove all his
life to construct.
I have just written the wo rd system . For it is in fact the tireless quest for
a vision of things and men at once all-embracing and unified, in short th e
search for an all-encompassing coherence, that is th e very hallmark of Pla-
to's thought. His theory of the relations between music and trance-for it
is a true theor y, as we shall see-i s consequently not separable either from
what he thought of music or trance or from his general theor y of the
world. In order to reach the chick of rhe matter, char is to sa y, these ver y re-
lations, we must begin from the srarr. I shall do my best, however, to limit
myself to the indispensible.

Mania and It s T erminology

That the word mania among the ancient Greeks me ant w h at I ha ve been
calling "trance " throughout thi s boo k is something char will become ap -
parent as this chapter proceeds. For the moment , therefore, let us rake this
fo r granted and begin by askin g ourselves \\"hat Plato's co nception of ma -
ma was . E ven so, one point sho uld still be clarified . ,\!th o ugh from our
perspecti ve, the \,·ord mania can be considered as th e ex act equivalent of
"tr an e, .. depending on rh e co ntext , I shall tr ansla te it so metimes by
'·madness" and so metime s by "fr en zy.·· The reader should see no inconsi s-
tency in this. ·'J\1adness '' and "frenzy'' are in fact more acc ur ate transla-
tions of mania as the Greeks under stoo d it, and it is their viewpoint char is
imp orta nt to u s at the moment. \Y/e should note, however, char although
"m adness " (o r rhe French /olie ) is, alo ng \,·irh "fr enz y," the most frequentl y
used tr ansla ti on of the Greek mania, chis has nor ah ,·ays been the case.
During the Ren aissance , to mention o nl y o ne period . it \\·as rendered in
French as fureur and in Italian as furore (Cicero sai d "furor" ), both of
which \Vere just as legitimate as Jolie, albeit signif ying another aspect of
things. T oday, Orlando fur ioso ought to be translated "Mad Orlando"
rather than "Furious Orl ando ." In the seventeenth and eighteenth centur -
ies , Michel Fouc ault points our in his Hist oire de la Jo lie a /'age classiqu e,
rhar /ureu r was a "technical term in jurisprudence and medicine" designat-
ing "a ver y precise form of madnes s."
Music and Trance among the Greeks 189
This said, what did the word mania mean for Plato' The answe r is to be
found partly in the Tima eus and partl y in the Phaedrus . ' In the Timaeus
(86b ), having given a general account of the diseases of the body (soma
nosemata ), he turns to those of the soul (psyche ),' whose specific disease,
he says, is dementia (anoia ), which is of two kinds: either madness (ma-
nia ) or ignorance (amathia ). And he adds : "Whatever affection a man suf-
fers from, if it involves either of these conditions, it must be termed 'dis -
ease'; and we must maintain that pleasures and pains in excess are the
greatest of the soul's diseases." Oddly enough, but significant no doubt ,
Plato's reflections in the Tima eus on the classification of mania stop there .
It contains no allusion whatsoever to something found in the Phaedrus, a
much earlier work, in which he makes what is for us a crucial di stinction
(265a-b ) between rwo different kinds of mania, one arising from human
diseases (nosematon anthropinon ) and the other from a divine state
(theias exallages ) "which releases us from the customary habits." He then
divides this stare int o four sons of mania, four different kinds of madness
each inspired (epipnoian ) by a God: "ma ntic," (by Apollo ), "relestic," (by
Dionysus ), "poetic" (by the Muses ), and "erotic" (by Eros and Aphrodite ).
This is how in a celebrated and often quoted p assage, Socrates summa -
rizes an d characterizes in one word the four forms of madness he had just
described in his discourse to Phaedrus on lo ve.
This manner of characterizing the fou r forms of madness merits closer
atte ntion. Three of the four are defined by means of epithets concerning
their function: the one leading to divination is called "mantic," the one
stirring the poet to creation is "poetic," and the one inspiring the lover to
transports of frenzied love is "erotic." The fourth, howe ver, is defined
from a quire different standpoint, a formal rather than a functional one . Ir
is called "telestic," meaning it entails teletai, or rites . In other words it is
"ritual" madness. As we shall see before long, this telesric frenzy is nothing
other than what we have been calling possession trance . For Plato, then,
that which characterizes the possession trance first and foremost and dis -
tinguishes it from other forms of tr ance, is the fact char it is ritual. About
these rites themselves he pro vides few details, no doubt considering it
pointless to elaborate on facts with which the reader of his day would be
familiar, and presuming chat the latter would know them well enough.
Fortunately, he nevertheless w rites enough about them (Laws 79la - b ) for
us to know the y essentially consisted of sacrif ices, dances, and music . We
can also be sure, moreover, whatever the arguments aroused by the inter -
pretation of the word teletai (we will return to it ), that these rites also com-
prised an initiatory aspect, of va rying importance, an d that the y must
have been secret, to some extent at least.
Now char we have defined each of the four forms of mania, and in par-
ticular relestic mania (possession tr ance ), let us return ro what the y have in
common. Di vine madness is the lot assigned to certain humans by the
 190 CHAPTERFIV

gods; it is thei r personal an d divinely ordained fare (theia moira ) and at tr

same time a manifestation of "enthus iasm" (enthusiasmos ),' wh ich is t
say t he presence of a god in the pe rson who is prey to t his frenzy . Tha t pe
son is in fact entheos or "engodded," as we ought to be able to say in ord <
to convey char "t he god is in him. ''4 T his is the wo rd tha t Plato utili zes t
designate the Sybil prey to mant ic madness in the Phaedrus (244b ), d
lover prey to mantic madness in the Symposium (179a ), and the poet po
sessed by the Muse in the Ion (534b ). It is also the word used by Eu ripid,
in Hippolycus to speak of the state of Phaedra, ' who suffers from a hidd e
ill (krypto pathei ) and who wants to put an end to her life : "O yo un g wo n
an are you not engodded [entheos] by Pan or Hecate, led ast r ay [phoita
by the awful Co rybantes, or by the Mother who reigns over the hill :
Would it be some faults [amplakiais ] committed against the ardent hun
ress Dictynna 6 for not having sacrificed victims for her?" asks the Choru
(We will come back to this fault later on. )
Bur let us return to Plato and the Ion. In this dialogue Socrates pokes fr
at a rhapsodise named Ion who is utterl y astonished to learn from him rh
when he recites the Illiad, he is possessed by Homer (Jon, 536b ). Here ti
words "enthusiasm" and "enthusiastic'' alternate with "possession" (kat ,
koche ), which stems from the verb kat echo , meaning "to hold firml y, r
rain, contain, detain, possess, rake hold of, occupy a place," in the m i:
tary sense of the word, and from there, "to possess," in the sense we use
here. When this passage of the Ion deals with poetic madness, enthusias
and possession are synon ymous. When he is inspired, the poet is said to I
... enth eos ("enthusiastic" ), katechom enos ("possessed" ), or mainomen<
~, ("enmaddened" ), a word not only reser ved for humans bur also appl icab
' to the gods, as is attested by the appellations of both Dionysus and H er
d es Mainomenos.
Other Greek words also serve to signify the state of possession .' In ti
Ion (536a - b ), speaking of the relation existing between the poets and ti
divinity inspi ring them and describing it under the fo rm of a "migh
cha in" hanging down, Socrates says : "One poet is suspended from 01
Muse, another from another; we call it being possessed, but the fact
much the same since he is held." And he adds that some poets "are fillc
with inspiration" by Orpheus, others by Musaeus. "But the majority a
possessed and held [katechontai te kai echontai] by H omer." Alongsi,
the verb echo, carry, lead, gr asp, hold," wh ich also fo rm s part of the ve
katech ein, 1 which we have already mentioned and which is probably me
frequently used with reference to possession, the Greeks also used la,
bano "take in one 's hands, grasp, rake possession of, occupy," and hen
"to take hold of, possess" in the sense we use it in this book. T his term
arrested in one of the oldest Greek rexes we have on possession, that I
Herodotus (consequently, one century before Plato ), in which we read,
Music and T ranee among rhe Greeks 19 1
the famous passage recounting how the Scythian Skyles was p unishe d by
his own people for having been initiated into Greek possession rites: "You
mock us, Scythians, because we give ourselves up to Bacchic tra n sports
and the god takes hold [lambanei] of us ... " (Herodotus, 4. 79 ). To con -
vey the fact that the possession state has become total, that it is at its
height, Euripides writes that the god "in his fullness floods" the human
body (som ' el the pol ys ) (Bacchae 300 ) of the one he possesses. And the lat-
ter, he notes, is then able to "tell the future," an observation tied to what I
said earlier concerning the transition linking possession and mediumship.
Among the Greeks, as elsewhere, there is no break in continuity between
the two, but onl y complete possession leads to mantic states, in other
w ords the kind of possession that presupposes a long experience of
trances and that is for this reason characteristic of officiants and not neo-
ph ytes.
The term th eol ept os, "seized by the god," which expresses the idea that
the god is present in the person subject (entheos ) to frenz y and that the lat -
ter ha s been invested or seized by this god, is more recent and does not oc -
cur under this form in Plato. But in the Phaedrus (238c ) Plato has Socrates
say, in a joking mood, that he would nearl y feel like a "nympholept," that
is to say possessed by a nymph.
Alongside the words m ania ("madness" ), enthusiasm os ("enthusiasm" ),
and kato k oche ("possession "), Plato uses , albeit less frequentl y, another
term also important for our purposes : epipn oia, "inspiration," which de -
rives from pneuma, "breath." He in fact uses this word in the famous pas -
sage from the Pha edrus alread y mentioned (265b ) in order to describe the
four m ani a : each is an ep ip noi a, an "inspiration. " The effect of epipnoia is
ro put the subject "our of his senses " (ekp h ro n ). Thus Plato says of the
poet th at he is "never able to compose until he has become 'engodded '
[entheos], and is beside himself [ekph ron] and reason is no longer in him"
(Ion 534b ).
Again, of the Delphic Pythoness and the priestesses of Dodona, he
writes that "when they were mad [man eisai] the y conferred great benefits
on Hellas, both in public and private life, but when the y were within their
senses [soph rono usai ] few or none" (Phaedru s 244b ). In the same passage
o f the Phaedrus, the opposition of "out of" and "within " one 's senses also
occurs a few lines earlier, to ch aracter ize the madness of love. And in the
L aws (790e - 791b ), Plato refers to this same loss of reason, when instead
o f poetic, mantic , or erotic m ania, he speaks of telestic mania, or posses-
sion trance: the Bacchantes, he says, are ekphron es or "out of thei r
senses," and it is the combined action of music and dance that resto res
them to their senses, so that the y are emphrones.
Thu s the four kinds of m ania- that is, of trance - all ha ve the effect of
thro w ing the affected pers on into a state of unreas on. This madness , how -
192 CHAPTER FIVE

ever, man ifests itself in very diffe rent ways depending on which mania is at
work. Neither the poet in his sta te of inspiration nor the love r p re y to m ad
love is ever described as being in the state of frenz y that manifes t s itself by
foaming at the mouth and rolled -back eyes, as is the case with Agave when
she is possessed by Dionysus -Bacchus and is about to dismembe r her son
Pentheus (Bacchae; 1122 ). Neither the poet, the lov er, nor the Pythoness is
ever depicted in the attitude of crisis char ac teri zed by a flung -back head
and an entirel y arched body, so often seen in the possession scenes pa inted
on va ses. None is described as undergoing the effects of music , or as hav -
ing dance as his or her princip al activity. All these man ifesta t ions are
symptomatic of telestic mania , which expresses itself through a set of par-
ticula r beha vior patterns totally differen t from chose produced by the ot h-
er kinds of mania.
The amnesi a suffered by a possessee up o n emerging from his trance is
also one of rhe character istic features of relestic mania . Witness the dra -
m atic account of Aga ve's horrified surprise w hen she learns, from the lips
of her father Cadmus, chat in her murderous fren zy, ha ving taken he r son
Pent heus for a wi ld animal, she had dismembered him and w renched off
his head. But what di stinguishes te lestic mania most rad ica lly from the
ocher three is th at it is the o ut co me of an offe nse, of misconduct cow ard
the god, w ho then m anifests his anger by striking the guilt y pe r son with
mad ne ss . The same could neve r be sa id of the ocher three mania. T h at it is
characteristic of telestic madness is form ally confirmed by Euripide 's text
cited above : "Would it be some offense commi tted aga inst the ardent
huntress Dicr yn n a for not having sacrificed a victim for her, " the Chorus
asks the "engodded" (entheos ) Phaedra. This is also attested by Plato' s
\\·ell-known text , which we will now examine at length .

Telestic Mania (Ph aedrus 244d - e)

In the Phae drus , as we saw ea rlier , Plato distinguishes between four kinds
of mania. Before prop os ing a classification (265a - b ), he stares by explain -
ing what each kind con sises of (244 - 49 ). Following a brief exordium cele -
brating divine mania in general, he tre ats a first type o f trance (mania ), the
one that seizes the "prophetesses" and chat he w ill lacer call mantic - let us
say divin aro ry tr ance. He then comes co the one that he will term "telest ic"
trance in his classific at ion. It is chis passage chat inte rests us. T his text ,
composed of a few lines (Phaedrus 244d - e), difficult co understand, and
well-known for its ver y obscurity, has been the obje ct of a ten -page study
by Linforth (1946b ), w ho before pro viding hi s own interpretat ion , sum -
m arizes chat of his predece sso rs, all of whom, he says, find it "baffling."
Wilamowitz -Moellendorf ( 1920, 1 :411 ) himself admitted that "he ne ver
found any explanation" for this pa ssage, which left him "pe rplexed"
Music and T ranee among the Greeks 193
("Eine Erkl arung habe ich nirgend gefunden und bin selbst ratios" ). To my
knowledge, no translation or interpret ation • of it , including that of Lin-
forth, is satisfactor y. The most recent tr ansl ation (1983 ), proposed not by
a He lleni st but by a well -known ethnopsychiatrist, G. Devereux , in his ar-
ticle "La crise initiatique du chaman chez Platon (Phedre 244d - e)" is
scarcely more convincing. But it seems th at in light of what was sa id about
possession in the first part of this book, a new readi ng of this text is p oss i-
ble . As presumptuous as it might seem, I will therefore chance a new trans -
lation of it here. ' 0 The read er will forgive me for prov icling one so close to
the Greek that it hardly resembles English, but I deemed it important to
follow Plato's train of thought as close ly as possibie.
The following is what Socrates sa ys: 11
Besides [1], indeed from these very diseases [noson ] and great woes
[ponon] rhar certainl y or igin ate [ek] in ancient offenses [menimaton]
committed by someone in some lineages [genon], [2) trance [mania],
[the very one] that comes upon those who [usuall y] indulge in ir by
giving them prophetic power [prophereusasa ], secures [eureto ]
deliverance [apallagen] through rhe use of prayers ro gods and of cul rs
[celebr ated in their honor ]; [3] thu s [othen], resulting indeed in
purifications [katha rmon ] and rites [teleton ], it [trance ) brings recover y
[exant e epoiese] ro the one who welcomes it [echonta ] for the present as
well as for rime ro come, [4) securing [eurom ene] for him who was
correctl y entranced and possessed [orthos man enti te kai kataschomeno]
release [lysin] from [his ] troubles.

Wha t does the substance of this text signif y ? Let us break it down ito its
four pans in o rder and br iefly analyze their contents:
[l ]: People suffer from diseases and woes that are the consequences of
certain offenses.
[2 ]: Di vinator y tr ance delivers them ch rough recourse to the gods.
[3] : Thus, purifications and rites bring them reco very.
[4 ]: He who is correctl y entranced is released from h is troubles.
As we can see, Plato 's expose consists of a serie s of facts th at are in di-
rect relation to tr ance and t hat succeed one another in a certain o rder. This
order is exactl y the one followed in the first p art of this book in order to
speak of the dynamics of possession. Broadly speaking, we are therefo re
de al ing in both cases w ith the same facts. Seen in this way, things become
clear. Socrates' discourse, nevertheless, is not. We must therefo re now en -
ter into detail and justify our analysis .
[1] What are those "diseases," "woes," and "offe nses" ? Plato does not
say, but the general context and situation to which he implicitl y refers do
not lea ve any room for doubt. The "tel estic " trance (mania ) with whic h he
is de aling here is the one he will later say is "inspired" by Dion ysus. On the
other hand, as we saw (above, 269 ), for Plato mania, which is a disease, is
 194 CHAPTER FIVE

of two kinds, on e due to "human diseases," the other to a "d ivine stare."
This being said, how can we concretely concei ve the relations th at Greeks
of his d ay discerned between Di onysus, the divine stare of mania, and dis -·
ease? One of the best-known tragedies by Eur ipides, Bacchae, whose topic,
as we know, is Di onysiac trance, pro vides the answe r to our question .
Di onysus is depicted as t ak ing terrible vengeance agains t th ose in hi s own
lineage who insult his mother's memory, refuse to celebrate his cult , and
chase our his followers. Finally, Agave (one of his mother's sisters ), struck
by mania, dis member s alive her own son Pentheus, also m ade mad by
Dionysus. Offense, madness (here, madness crisi s, in other words, dis -
ease ), woe (what more horrib le woe for a morher? )- rhis perfectly illus -
tr ates Plato's text, with the exception char in the Bacchae Agave's murder -
ous frenzy is provoked nor by "ancient" faults, char is to say, chose
com mitted by others long before he r, bur by her very own.
It is thi s reference to the ancient aspec t of the fault that has frustrated
the majority of commentators. 1 2 The difficulty is easily solved if we recall
that Plato considered mania crisis as resulting from fright, and fright as
rhe sign of a "defective disposition," or to put it m ore simpl y a certain
"weak nes s of rhe soul." He express ly scares chis in the Laws (790 - 91 ); we
will return to the subject lacer. The present text of the Phaedrus tells us
that Plato saw in this weakness of the soul, which predisposed one or an -
other indi vidu al to tr ance , the consequence of an ancient fault comm itted
in the lineage; in other wo rds, he cons idered it to be hereditar y, at least in
part. T he difference between the concrete example of tr ance de scribe d by
Euripides in the Bacc hae (fau lt attributable to the person himse lf) and the
abs tr act tableau drawn by Plato in rhe Phaedrus (fault attribut able to the
lineage ) is due to the fact th at Euripides refer s to legenda ry time , when the
cult of Dion ysus was in the m ak ing , 1 3 whereas Plato refer s to historical
time, w hen this cult had already long been institu ted. Let us say met a-
phorically th ar in chis p assag e of the Phaedrus, it is not Aga ve that Plato
has in mind , bur her descendants.
In sum, rhe first part of the present text describes a situation corre -
sponding to the one seen so ofte n in the first half of this book as being at
the o rigin of possession, that is to say, "histories of illne ss" th at result in
"cr ises ." As for the Baccha e, the crisis rh ar affe ct s Agave is p articularl y
tr agic, to be sure, bur Greek antiqu ity offers numerous co mp arab le exam-
ples.
[2] The second part of the text poses the problem of knowing just exact ly
who the people are whom Plato designates -r at her enigma tic ally, we
should say - as "usuall y indulging in tr ance ," which gives chem "prophetic
power." Without exception - at leas e to my knowledge - translators and
commentators alike have understood char thf people concerned were the
same as chose mentioned in the previous ph rase; in oche r words, chose suf -

"---
Music and Trance among the Greeks 195
fering from the woes of mania . Therein lies the error, for this is not the
case, for two reasons.
Fir st, in his general discussion of mania, which almost immediately p re-
cedes our passage, Plato begins precisel y by speaking of prophetic trance,
an d he most clearly indicates those whom it affects, giving as examples the
Delphic oracle and th at of Dodona. Consequently, the people he has in
mind a re the di viners. But here he refers to the diviners in general and not
necessarily to the two just mentioned , for as Dodds, who is evoking Plato,
tells us: "No r was prophetic possession confined to official oracles. No t
only we re legendary figures like Cassandra, Bakis, and the Sibyl believe d
to have prophesied in a state of possession, but Plato refers frequently to
inspired prophets as a familiar contemporary type" ( 19 51, 71 ).
The second reason for considering that mania having "prophetic power"
concerns diviners and not the diseased springs from the very logic of the
system, which would ha ve it that everywhere, and not only in Greece, the
diseased come to consult diviners and ask them what to do. In this respect,
Boyance ( 1936, 66 ) cites a case that perfectly illustrates the rule and con -
cerns Athens itself:

The legend that concerns the introduction of a cult or iginating in

Eleurheria conforms to the habitua l scheme: A fault; the Athenians
refuse to welcome with honor the statue of the God brought to them by
a certain Pegasus; the result is a disease that in this case affects not the
spirit but the body ... ; recourse to a prophetic and inspired authority,
this time as so ofcen before , the Delphic Oracle. The consequence in the
institution of the cult and the healing of the disease .''

Exactly wha t role did these diviners have? Among the Greeks as else-
where, it is clear that the y were first aske d to identify the deity responsible
for the woes afflic ting the diseased person in order to know to whom pray-
ers should be addressed, a preliminar y requisite to any hope for healing. 15
(3 ] After the sick persons have sought consultation, the diviners, having
expressed their opinion, prescribe purifications (katharmoi ) and rites
(te letai ). What exactly we re these teletai so characteristic of this mania
that Pl ato ,,·ould call it "telestike"? Or, if one prefers, what were these rites
so characteristic of this trance that Plato would call it "ritu al?" In the
Bacchae, rites are frequentl y mentioned. In the beginning verses of the
traged y, Dion ysus presents himself and says : "To this of Hellene cities first
I come, having established in far lands m y dances and rites , to be_ God
manifest to men " (Bacchae 20- 23 ).
The wo rd telete is frequently translated as "initi ation," but Boyance
(1932, 42 ) and Dodds (1960, 75 - 76 ), who both take a great interest in
telestike, ag ree that the word generally des ignates the ritual of mystery
cults. Prayers, sacrifices, purifications, dances, and songs are thus part of
196 CHAPTER FIVE

the telete. But Boya nce, who examines the se rites in the Orphic context, in
which the y pla y an important role , emphasizes that the y are expected to
be effecti ve. The wo rd teletai, he w rite s (193 2, 48 ), seems ro designate
"rites w hose thrust is ro more or les s force the deit y, rites whose most im-
portant p art is recourse ro effecti ve formulas ." He tells us elsewhere (ibid.,
4 7, 161 ) that the y are chanted form ula s. But force the deities ro do w hat?
To m anifest their presence . And Boyance asks: "W hat exactl y was this
pre sence which, ro our way of thinking , is the main goa l of these rites th at
are ma inly voc al and b aptized by the name of teletai?" (ibid., 54 ). He an -
swers: "We can only think about what the science of teleta i, wh ich became
telestike, is for the Neoplaronists . It essentiall y cons ists in provoking the
arrival of the deity, but this operation is a very material one: the god ani-
mates the statue representing him and telestike is the art of consecrating
statues." It is no t unreasonable ro think th at in Di onysiac cults the wo rd
teletai also ser ved ro design ate rit es aim ing at "pro vo king th e arri va l of the
deit y," bur thi s rime in a different way : by arous ing possession trance. For
Di onys iani sm , teletai could thus signif y, in part at leas t , the art o f conse-
crating nor st atues but initiates. We cannot sta re it definitely, but we can at
least hypothesize it.
Whatever t he case m ay be, rites necessaril y imply acco mpli shing a series
of acti o n s in a certain order, respe ctin g the prescribed o rder, correctly fol-
lowing a certain trad iti onal behavio r. Thi s aspect of things must abo ve all
be retained here ; we shall soon see why .
Let us now turn ro th e second pan of (3). The problem is ro know exac tly
who is the one "welcoming" this tra nce and what aspec t of tr an ce , pro -
phetic or tel esric, is involved.
Evidently, this trance that "b rings recove ry" is the one ro w hich Plaro
actually refers, in other wo rds, "relesric" tr ance . As for the one ro w hom
this trance "br ings recovery, " he can only be th e one w ho needed it , that is
ro say, o ne of the people referred ro in [1). This sick person on ly rec ov ers,
the text says, if he "w elcomes this trance for the present as well as for rime
ro come." Wha t does thi s mean? Ob vio usly, that the person concerned
will have ro, from then on, regularly comply wi th the ritu als of relesric
trance. In other wo rds, he w ill recover only if he has be en initiated an d in -
tegrated into the cult of th e god responsible for his woes; in brief , on ly
after he has become member of what th e Greeks called a thiasos, let us sa y
a "college of the possessed. "
(4) We ha ve just seen that the o ne who "w elcomes" relesr ic trance "re-
co vers." But how is he "released from his troubles?" The last p art of the
text sta res this clearl y : by being "correctly entranced and possessed, " in
other words, by co rre ctly (orthos ) obser ving the rites of the trance; in
shore, by pr acticin g ritual trance . T his is uncieni ably ho w this phrase,
which has frequentl y been badly tr ansl ated , must be undersrood. Let us
Music and Trance among the Greeks 197
emphasize here that if, in his last sentence, Plato speaks of a correct way of
being entranced, he also implies there is an incorrect one. Otherw ise his
text would be meaningless. For Plato, the re th us existed two oppose d as-
pects of trance, one ritualized and the other nor. Among the Greeks, we
once again find the opposition of ritualized versus nonritualized tra nce,
which, as we saw in the first part of this book, is one of the universals of
possession.
1ow th at we h ave completed our analys is, the scheme underlying Pla-
to's text can be rewritten in a different way, one that will more clearly
identify the nature of telestic trance. The episodes in the rexr succeed one
another in the following manner: (1] diseases and woes (i.e., nonritual -
ized trance crises ) due to the offended god's anger; [2] consu ltation of the
diviners, who through prophetic trance identi fy the god in question and
prescribe the adeq u ate prayers and cults needed to "secure deliverance" of
the sick person; [3] purification and rites, which imply initiation and inte-
gration of the sick person inro a "college of the possessed"; (4 ] practice of
ritu al ized possession trance, healing.
Before going further and leaving this highl y instructive text, let us ob-
serve that the obscurity of this pas sage of the Phaedrus (244d - e) which has
caused so much ink to flow is due ultimately to the fact that Plato, who
will define four types of mania much later, here describes only its different
aspects . And at this stage, which is that of divine mania in general, he does
not distinguish between them. From one end of the text to the other, mania is
the grammatical subject, but Plato does not specify the form of mania in-
volved. It is up to the reader to know that in [1] he speaks of telestic tr ance
in its nonrirualized aspect, in [2] of trance in its divinatory aspect, in [3] of
both aspects one after the other, first of divinatory trance wh ich prescribes
the rites, second of telestic trance which brings rec ove ry, and lastly in [4 ]
of ritualized telestic tr ance .
For Plato, tele stic trance thus appears to be profoundl y ambivalent. It is
at one and the same rime sickness and healing of sickness, which should
not surprise us. Let us say, using Zempleni's term, that its main dimension
is therapeutic. But among the Greeks this "therapeutic dimension" could
ver y well be a later aspect of po session trance since the latter did not al-
ways ha ve rhe amb ivalent character, both divine and shameful, that it had
in rhe days of Plato. He himself tells us so, srill in the Phaedrus (244b ),
where we can re ad : "madness was accounted no shame or disgrace by the
men of old who gave things their names ."
Ir is clear that if Plato feels obliged to write that in olden times mania
was "accounted no shame nor disgrace," this is because in his time it was,
if not by eve ryo ne at least by some, and not wi th out conflict as the very
subject of the Ba cchae tragedy demonstrates. Inde ed, Plato himself seems
to h av e had ambivalent feelings, if nor toward mania - "madness" in gen-
198 CH APTER FIVE

eral - at last coward its telestic form; in othe r wo rd s, coward possession

trance. The se feelings expla in , at least in part, the conception he had oi
the rel ations existing between this form of trance and music and dance. It
is therefore important to know them.
In the Phaedrus, which, as w e know, is the dialogue in wh ich he sets
forth his conception of love, Pia co precedes Socrates' discourse on rhe four
forms of di vine m ad nes s (mania ) w ith a shatter ing preliminar y stateme nt.
Ha ving said that , contr ary co what cert ain people think, the belov ed
would be wro ng co prefer the lover "in a calm state of mind" (sophronei ) to
the one "made m ad" (ma ineta i) by love under the pretext that madness is
an evil , Socrates declares: "In realit y, ou r gre atest blessings come co us by
way of madness [manias ] w hich indeed is a divine gift" (Phaedrus 244a ).
Here we are back co the conception of m adnes s as a "divine lot" (theai
mora ) resulting from a "div ine fo rce " (theia dynamis ), an idea we alre ady
encountered in the Ion (5534c ). A fine "pr aise of folly," as Era sm us wo uld
ha ve put it 1
But is it sincere? D odds ( 19 51, 64 ) sees it as a "conscious paradox ," a
delibe rate p rovoca tion th at "st artled the fo urt h-cent ur y Atheni an re ader
hardly less tha n it srartles us ." Bur matters may nor be as simple as th at. If
we look m o re closely, we quickl y see that everyth ing depends upon the
kind of madness we are talk ing a bou t. Socrates, as we saw, distingui shed
betwe en four different kinds . Ever ythin g indicates th at his eulogy of
"ero tic m adne ss" is perfectl y and unre ser vedly sinc ere. "Of all forms of en -
thusiasm," he tells us, it is "the best." Th rough it, when one "sees be aut y
on earth, remembering true beaut y, one !eels one 's w ings growi ng"
(Phaedrus 24 9d - e). For Placo, indeed, the madness of lov e is the di rect re-
sult of contemplating beauty, and beauty -t hat of the yo ung bel ove d- is
the best access to the good and the tru e, which is to say rh ar which is most
important in the world, the essential, in the strict sense of the wo rd. Ir is
thus easy to understand why Plato conside rs mad love - o r erotic m ad ness
as he calls it- as a gift from the gods, as an inspiration , as an enthusiasm,
and why he shou ld present it as such . I per so n ally do not see any par adox
in this at al l.
Thin gs are less clearcut , however, as soon as we move from erotic mad-
ness to poetic m ad nes s. Here, the least o ne can say is that Socr ate s is
laughin g up his sleeve as he poke s fun at the unfortuna te Io n. W hen he
tell s him, w ith utm ost serious ness, that it is Homer who po ssesses him, he
asks h im: "When yo u recite epic verses correctl y and thrill the spect ator s
mos t deepl y .. . are yo u not beside you rself, is yo ur soul not tr anspo rted
by enthu sias m? " (Jon 535b - c). And Ion ack n owledges that something of
the sort does happen w ithin him: "W hene ver I recite a tale of pit y, my eyes
are filled with tears, and whe n it is one of horror o r d ismay , my hair stands
up on end with fear, ' 6 and my he art goes leaping" (ibid ., 535c ). This is like
Music and Tr ance among the Greeks 199
the shi ve r running up the spine of the officer-cade t hearing his national an-
th em, or of the veteran he arin g the Last Po st . So Socrates is in fact describ-
ing something real, so methin g th at ex ists, and is there fore only half-joking.
Indeed, elsewhe re (Phaed rus 245a ), aga in wit h reference to p ossess ion
and th e madness of the Muses (M ouson katokoche te kai mania ), he says:
"Bu t if any man comes t0 the gates of poetry wi th out the madness of the
Muses [manias Mauson], persu aded that skill alone will make him a good
poet, then shall he and his works of sanity w ith him be brought to nought
by the poetr y of madnes s, and behold t he ir place is nowhere to be fou nd ."
Here again, I see nothing paradoxical. Music - or rather "mu sic tech -
nique" (mousike techne ) for "m usic," meaning "of th e muse s," was origi-
nall y an adjective - being anothe r aspect of poetry, is also "a science of
love matters" (erotikon episteme ), w ith respect to "harmony and rhythm"
(Symposium 187c ). Both music and poetry therefore proceed in part from
love, and it is thus logical that both erotic and p oet ic m adness should be
almost equally dear to Plato's heart. It is still true , ho wever, that he ne ver
jokes abou t the first whereas he does on occasion abour the second. Ah,
one ma y say, bur Ion isn't reall y a poet, "a creator," he is just a rh apso dise
and this is why he can be ridiculed. T he fac t remains , howe ver, that both
possession and enthusias m are being mocked throu gh him, and wheth er
one likes it or nor, Socra tes is no longer comple tel y serious here.
As for m antic mania, that is to sa y, di vin ato ry frenzy, Plato see ms to
hold rathe r the same op ini on of it as of poetic frenz y. Ir is hardl y necessar y
to say much more than that. Let us simply say that on the whole he be -
lieves in it. Nor that he holds the prophetesses of Delphi or Dodona in a ny
particular esteem. Inde ed, he claims that w hile "they we re mad the y con-
ferred great benefits on He llas," bur immediatel y thereafter he adds th at
while t hey were "in their senses" the y conferred "few or none" (Phaedrus
244b ).
Things are quire different, o n the other hand, when Plato deals wi th
relesric mania, in other wo rd s, with possession tr ance. Once and only
once is it qualified as "inspir ation ," and this is because it is part of divine
mania in general. Ir is nor even once called "enthu siasm ." Ir is "posses -
sion ." This is rhe wo r d Plato consranrly uses to speak of it , and chat's all.
This is al re ady sig nific ant in and of itself, but there is more. To w hat ex -
actly is this mania, which like all divine mania is a gift of God, tied? T o
fright, Plato sa ys in La ws (790e ), a fright resulting from a certa in "weak -
ness of rhe soul" and that is rhe "practice of cowardice " (ibid . 791 b ). Plato
puts chose who suff er from this fright in the sa me group as cr ying children
rocked to sleep by their nann y- a com p arison hardl y flattering for the
Bacchantes . How then shall we class ify those Bacchic dances wh ich
"mime drunken people" and whose rh ythms (ry thm oi ) are equally "a ppro-
priate to illiberality, and insolence or madness or other evils" (R ep ublic
200 CHAPTER FIVE

400b )? Well, it would be best to "separate it off both from p acific an d

from warl ike dancing, and to pron o unce that thi s kind of dancing is unfit
for our citize ns" (Laws 87 5c-d ). Plato thus did not hold these bacch ic·
dances in great esteem."
Thus, when Plato sa ys th at mania is for us "the greatest bless ing and a
gift from God," there is indeed "a conscious paradox." In making this
statement, he since rely has in mind the three mania, erotic , poetic, and
mantic; as for relestic mania, his feelings are certainly mixed. 18
Plato uses rwo words to designate th~ possessed, or, more gene r ally,
adepts of possession cults: "Bacchantes" (bacchoi ), which of co urse is de -
rived from Bacchus, and "Corybanters" (korybantiontes ), or th ose who
act as Corybantes, the two words being more or less synony mou s as he
uses them. Oddly, however, he never uses the word maenads (mainades,
from mania )- rhar is to say, rh e mad, frenzied, or craze d female compan-
ions of Dion ysus so often depicted in p aintin gs and symboliz in g the tr an ce
state in Greek religion - eve n though the y occupied the place they did in
anc ient Greek literatu re. Why?
Perhaps because the word mainades designates very particularly, it
would seem, Dionysus ' permanent companions who had come with him
from Lydia (Bacchae 55 ), whereas bacchae, wh ich also designa te s them,
could be indiscriminately app lied to all female adep ts of his cult - or at
least this is what eme rges from a reading of Euripides. Th e wo rd s "Ba c-
chant" and baccheuein ("to behave lik e a Bacchant" ) we re nor onl y re-
served for Dionysus -Bacchus; they were of general use and could as w ell
apply to the worship of anothe r daimon, in particular to the Corybantes 19
and to adep ts of Cybele wo rship . If Plato abs tains from any reference to
the m aenads, it is perhaps because the y have no connect io n- or in an y
case have much less connection th an Bacch ants or Bacchantes - w irh the
therapeutic dimension of tr ance. To speak of "Bacchantes" or of "those
\\·ho beha\·e like Corybantes" (korybantiontes ), an d thus implicitl y to re -
fer to the cult of Cybele, "Mo ther of the Gods," is, on the contra ry, to refer
to a form of possession whose ther apeut ic dimension is primordial.
Dodds points our that the worship of Cybele sho uld , in fact, be viewed
from this angle. Having obse rved ( 1951, 77 ) that or her deities besides
Dionysus could "cause mental trouble" and that "presumab ly all could
cure \\·hat they had caused if their anger were suitably appeased," he tells
us char "b,· the fifth century the Corybantes at any rare had developed a
speci al ritual for the treatment of madness. The Mocher, it wou ld appea r,
had done like\\·ise (if indeed her cult \\·as at that rime distinct from rh ar of
rhe Con·bantes )." In Th e \\.'asps (119-20 ), \\·hat is Aristoph -anes' point of
reference when he \\·ants a typical situation to express a son's efforts to
cure his father, Bdelycleon, of his mania for judging) Qu ire naturall y, rhe
Corybantes. The son makes his father '"corybantize," bur the latter res.isrs
Mus ic and Tr ance among the Greek s 201
and " runs off with the drum ." What does Plato do, also with a n int ent co
mock, whe n he wanes to characterize Phaedru s' mania for speec hi fyi n g ?
He depicts him as delightedl y "syncorybancizing" w ith ano ther pe rso n
"whose sickness [nosounti] is wanting to listen to speeches" (Phaedrus
22 8b - c). All chi s sufficiencly demon stra tes, I cruse, how far "cor y ba nt iz-
ing " an d be ing affl icted b y mania - as well as being cured of it- we re, in
Plato' s rime, synon ymous.
ow th at we h ave dealt w ith the \\ ·ords, we can come to what Pl ato
thou gh t about the relation of d a nce and music to trance .

Music and T ranee: Pl ato's Theor y

Pl ato ne ve r gathered together his ide as a bout the rel ati on of tr ance to mu -
sic and dance into an organized a nd , as it were, a utonom o us whole. T he
fa ce remains, howe ver, cha r these ideas do form a coherent whole char is
nothin g less than a theory. This becomes quite app a rent if one ca kes the
troub le to collate a nd logically articul a te a number of passages dealing
w it h teles tic m a dne ss chat o ccur in the Phaedrus (something we have al-
re ady seen at length ), in the Laws, an d in the Ion . ' 0 Thi s is wha t I n ow pro -
pose to d o, and the be st scarring point is a passage from the La ws (790 d -
79 1 b ), consisting o f a bout cwenc y lines, w hich sho uld be quoted in full ."
De ali ng with the exercizes chat sh o uld be used to form th e souls o f ver y
yo ung boys, Pl ato examines the meth ods of ra ising in fa nt s, a nd , in chis
context, ad voca te s the virtue s of continuous m ove m ent , w hich is "in all
cases sal ut ary" bur especially in the case of th e new born "because it is as if
they we re always n av igating. " Experience, he says, has brought home the
adva nt ages of this method co "chose who nurs e sm all children" as well
as - an d here we come to the passage char int eres ts us -
to rhe women who rirualize [telousai ] in rhe healing [iamata] of rhe
Coryba ntes . For \\·hen mothers have children \\·ho suffer from insom nia
and \,·ant to go to sleep, lull them ro resr , rhey bring them nor srillness,
bur chis ver\' mo\ ·ement [kinesin] , for they rock [seiousai] them
ceaselessly in rheir arms [rhey bring them) nor silence, bur melody
[melodian], [so char they ) "en -aulize'' [kata11/ousi] rhe children jusr as [is
done ro] Bacchante s our of rhe ir senses [ekphronon], by using rhis cure
[iaseis) [consrirured by such forms ] of movement [kineseos] as singing -
and -dancing [choreia ]" and "mus ia" [mouse] ."
Th e reason behind a ll chis, Pl ato goes o n , is char:
rhe sufferings [pathe] of both are, in brief, frights; frights [deimata] char
come from a defective disp osition [hexin phaulen ] of rhe soul [psyche ].
So whenever one applies an external shaking [seismos] to suffering s of
this kind, rhe external movement [kinesis] overpowers rhe internal
202 CHAP TE R FIVE

movement [kinesin] of fear [phoberan] and madness [maniken] and by

thus ove rpo wer ing it, it brings about a manifest calm in the soul and
a cessation of the grievous palpitations of the he art, which had existed in
each case. T hus, it produces very sat isfactory results . It (the
movement]" brings sleep to the ones (the newborns ] while to the others
[the Bacchantes ], set to dancing [orchoumenous] and listening to the
aulos [auloumenous], it brings wakefulness [and], wit h the help of the
gods, made favorable by happy auspices, an d to whom each one
offe rs sacrifices, it completes [for them ] the enjoyment [ekein] of a sound
state of mind [hexeis emphronas] instead of a state [diatheseon] (that
is] in our eyes madness [manikon] .

Just a wo rd first to justify two odd ities in this translatio n : the terms "en -
aulize" and "mu sia." Th e Greek kataulousi is always translated, w hether
in English o r in French, by words referring to song or speech (such as "to
cast a spell" ), whereas in fact it refers to a musical instrument, the aulos. 24
It is all the more important to respect the Greek text because the aulos, as
we shall see later on, is the instrument par excellence of trance amo ng the
Greeks, hence th e w ord "en -aulize ." As for rhe word mousa, the reason it
has been rendered here by "musia"-meaning here "thin g of the Muse" -
and not, contrary to customa ry usage, by "music" is simp ly because it does
nor actually mean "m usic " (cf. Bailly's dictio nary ), but ra ther "muse" first
and foremost, and hence "science," "art," or "song," or else "pe rsu asive
wo rd s." There is every reason to think that the tunes playe d on the aulos
du rin g possession cerem onies we re, as everywhere else, "mottoes," or
more precisel y "pe rsu asive wo rd s." And w hat could there be in common
betwee n the melod y made by nurses to the babies they are rocking - in
othe r words lull ab ies- and the aulos theme s pl ayed to the Bacchantes,
ocher than the fact chat both are "pe rsuasive words"? If mousa had to be
translated, in the present context, by an alre ady existing term , then "per-
suasive wo rds" is the one we ought to choose. Ne vertheles s, I prefer-still
in chis particular context of course-to use "musia," wh ich still stays very
close to the word's origin and does no t prejudge what exac tly Plato had in-
tended to say. But in any case, there is no reason to translate it by "music. "
Furthermore, to do so would be to misrepresent Plato, since whe re as he
constantly uses the word mousike, as we know, he abstains from doing so
here. To conclude, we need to remem ber th at in this passage there is no
mention, strictly speaking, of music. It will be readily ag reed that this is
important.
Movement, on the ocher hand, is rep eatedly mentioned. Indeed, it is al-
most the main subject matter of the passage since it is to movement that
song, dance, and "mus ia" ultimately lead, as Plato expressl y indicates
when he w rites: "this cure [constituted by such forms ] of mo vement as
choreia and mouse." Bur movement also has a cen tr al place in his general
Music and Trance among the Greeks 203
theory of health. It is thus very signific ant to note th at mo vement should
als o be at the very center of his theor y of the rel ation between music and
trance.
In the Timaeus, w hich as w e know contains an accou nt of his general
theory of the world system, humans included, Plato , after ha ving dealt
with human psychology, anatomy, and physiology then goes on to de-
scrib e the diseases to whi ch man is prone, those of the body first , then
those of the soul ( Timaeus 86g ), and it is here th at we find the definition of
madness and its two aspects ment ioned earlier. Imm ediatel y after that, he
cackles the problem of health (87c ). In order to maintain it, he observes,
noth ing is more important than to pre serve a correct relationship between
so ul an d body, and the me ans to this end is "never to mo ve [kinesin] the
soul without the body or the body wi thout the soul, so chat, each defend -
ing itself aga inst the other, the two sides will retain their balance and their
health" (88b ). He then gives two concrete examples: "It is ther efo re neces-
sary that the m athematician, and anyone w ho strenuously performs an y
intellectual activ ity, should also give movement [kinesin] to his bod y by
pra cticin g gymnastics. While the man w ho is diligent in molding his bod y
must in turn pro vide his soul with m ove ment by cultivating music
[mousike] and philosophy in general, if either is to deserve to be called
both fair and good" (88c ).
Moving on then from the particular to the general, he secs out his princi-
ple of the complementarit y- or the anta go nism - of internal mo vement
and external movement (T imaeus 88c - e):

T he vario us parts , likewise, must be treated in the same manner, in

imitation of the form of the UniYerse. For as the body is inflamed or
chilled ,,·ithin by the particles that enter it, and again is dried or
moistened by th ose without, and suffers the affectiom consequent on
both these move ments, whenever a man delive rs his body, in a sta te of
rest, co these movements, it is overpowered an d utterly perishes;
whereas if a man imitates th at ,,·hich we have calie'd the nurturer and
nurse of the Universe, and never, if possible, allows the body co be at
rest but keeps it mo ving, and by continually producing internal
vibrations, [he] defends it in nature 's way against the inward and
ourn·ard movements . ... [88d - e]

And he concludes: "W herefore the mo vement chat is best for purgings and
renovations of the body co nsists in gymnast ic exercises; and second-best is
the mo vement pro vided by swayi ng vehicles such as boats o r any co nvey-
ances chat produce no fatigue" (89a ).
All that, however, only concerns the "composite living creature an d the
bodily part of it" ( 89d ). But now we find that the same treatment should
be followed for our "t hree kinds of soul" (psyche ) th at "a re housed w ithin
204 CHAPTER FIVE

us in three regions" an d each of which "has its own movements" ( 89e).

"We must therefore see to it chat the y all have cheir movements relatively
co one another in due proportion." As for the "most lordl y kind" of our ·
so uls, which "raises us up from earth, " it coo must retain "its own congen -
ial. .. movement." The se movements (kineseis ) which have an affinity
with the "d ivine part within us .. . are the intellection and revolutions of
the Universe. These each one of us must follow," and when they are dis-
coned they muse be rectified "by learning the harmonies and revol utions
of che Universe" (90a - d ).
This summary of the Plaronic theory of movement as che basis of health
in man's body and soul makes it fa irl y clear that his ocher theor y, char of
movement as the b as is of Corybantic healing, is nothing ocher than a par -
ticular case, a particular appli cation, of the first. For the second does in
face contain the rwo broad principles of the first: on the one hand char of
rhythmic rocking, on the ocher chat of the ant agonism between the two
secs of movements, external and intern al, char counterbalance one an-
other and thus ree stablish equilibrium when it has been lose. The distur -
bances affecting chose afflicted by relesric mania spring from the effect of a
fright on a weak soul. This fright is an internal movement (kinesis ). Dance
and "musia" then intervene as an "external shaking" (seismos ), and by vir-
tue of their specific movement, which overc omes the earlier one and"over-
powers" it, resrore calm and tranquillity. Bue how exaccl y does this exter -
nal m ove ment ach ieve such effects' Plato does nor explain chis here. Bur
there is justification for chinking chat, in acco rd ance wi th terms of his gen -
eral theory sec forth in the T imaeus, they deri ve nor only from the virtue
sui generis of the rocking movement, but also from the face chat chis move-
ment is inseparable from the "ha rm onies and revol utions of the Un iverse ."
By virtue of th eir movement, da n ce and "musia" are thus able, in shore, co
reintegrate into the orde ring of the Universe the individual who had be-
come sepa r ated from it by che disordered mO\·ement of fright. To be sure,
Plato feels bound co add (bur giving the impression that for him it is just a
matter of form ) char the desired result does nor occur without the benevo-
lence of che god . In any case, ho,,·ever, the teletai (rices) are aimed precise-
ly coward this end. Such is che the o ry of Con·bantic cure put forward by
Plato in the Laws . "This is," he concludes, "co put ic sho rtl y- quite a plau-
sible account of the matter'' (78 1 b ).
Following chis text bearing upon the curati ve funct ion of trance, and
upon the central role played by movement and, therefore, abov e all, by
dance (since music plays a secondary role and more over is not even men-
tioned by th at name, we should stre ss ), lee us turn co another text , this
time one that relates dance to what we shall call, for the sake of conve-
nience, music, even though, here aga in , it is also not given char name. We
read in the Ion (536c ): Those who are corybant izing (korybantiontes )
Music and T ranee among rhe Greeks 205
"have ;i sharp ear for one tune [melo us] onl y, the one w hich belongs to the
go d by w hom the y are possessed [katechontai] and to that tune they re -
spo nd freel y wit h gesture [schemat on ] and speech, w hile the y ignore all
ochers ." What doe s this passage tell u s? In the first place chat the possesse d
person react s to one tune and one alone. W hy ? Wit h out doubt because it
is his. Taking adva ntag e of wha t we hav e found to be the cas e everyw here
else in these matters, I shall call thi s tune his motto . When he hears th is
musical mono, this tune, how doe s the "co ryb anti zer" re act? The text,
and even more so the context - which I hope I m ay be forg iven for not
quoting - indicate the answe r to this quite cle arly: he go es into trance. 2 5
And how does he manif est this tr ance? By finding "ges ture s and speec h "
enabling him "to respond fre ely" to this tune, 2 6 and, consequently, by
dancing. We are therefore we ll aware of the role pla yed by music in trig-
ger ing th at trance . Ir d oes not ope rate by means of so me sui gene ris virtue
or o th er that produces emotional effects of a specific kind ca pable of put -
ting people beside them selves. It is not as music - th e wo rd is mine and
does no t figure in Plato 's text - th at it acts, it is as a melody tha t has a
me a nin g to wh ich it is necessar y to resp ond ; in shon , it acts as a coded sig-
nal.
On the bas is of the d at a give n bv the t hr ee above -menti oned dialogues
(Phaedrus, Laws, and Ion ), Plato's the or y of the rel ations between tr ance
and mu sic ca n now be formulated as follows : "People who are psycho log i-
cally somewha t fragile, and who as the result of god's ange r suffer from di -
vine madness, cure themselves by p r acticing ritual tr ance, wh ich is trig -
gered by a mu sical motto and takes rhe fo rm of a dance; music and dance,
by th e effect of their m ove ment, reintegrate the sick person into the gen -
era l mo vement of the co smos, and this healing is brought about th anks to
the benevolence of gods who ha ve been rendered propitious by sac ri-
fices. " 11
Formula ted in this way, Plato's theorr on the rel atio n of trance to dance
and music appears ar first to be very different from the one usually set
forth; I shall later explain how . In rhe second pl ace, ir see ms to coincide
broadly w ith rhe th eo ry proposed in rhe first part of this book , except on
one essential point, which is rh ar of the ro le ass igned to possession itself in
rhe he al ing process.
Let us begin wirh what the rwo t heo ries -hav e in common. On the whol e
rhey coinci de with respect to rhe role allo tt ed to music. Music does pl ay a
pan, granted, bur iris n o more than one element among others, wh ich in-
clude ritual s, dance, m ovement, and, of course , the be.ne vo lence of the
gods . Music's role is so far from being predominant rhar the word mu sic
ap pe a rs only once in Plato's rexes . (I shall later return to Plato 's way of re-
ferr ing to it in the context of possession, sinc e ir is significa nt .) It is the
supp ort of d ance. Mus ic also triggers tra nce; bur never once is ir presented
 206 CHAPTER FIVE

as doing so by means of an emotional charge of an y kind. It triggers trance

because it is a motto, in othe r wo rds a signal.
Let us now come to the divergences. The central point of my theory is ·
that music, by acting as both the signal for trance and the support for
dance, essentially allows the possessed person to publicly identify himself
to the god possessing him or her. It is this process of public identification
with the god that is for me the mainspring of possession trance, and when
this possession has a "th erapeutic dimension" of any importance, it is this
public identification that is the principal agent of the cure. In Plato the
gods play a part, it is true, but only - and this barel y indicated - in order
to strike people with madness. At the healing stage their role is limited to
viewing things with benevolence. At the outset, when defining the four
m ania, Plato does mention "en thusi asm" and "inspiration," and mania is
presented as caused by the gods. Her e, however, in the texts concerned
solel y with telestic mania -p ossession trance - nothing of the sort is
evoked. Telescic mania appea rs to be emptied of its religious content. Seen
in terms of its therapeutic dimension, it appears as an alm os t purely pro-
fane and mechanical affair. If I had to draw a parallel between Plato's the-
or y and my own, I would say that, for him, the healing process is achiev ed
by reinsertion of the individual in the cosmos as a result of the movement
of dance and music, which ree stablishes harmony with the universe,
whereas for me healing is achieved by the reinsertion of the ind ividual in
societ y as a result of the movement of music and dance, wh ich provokes
identification with the god. The difference can be expressed as in figure
5 .1.
To sum up , Plato's theor y is both ph ysical and metaph ysical, whe reas
mine is both psychological and sociolo g cal; both are equally ph ysiologi -
cal.
The questi on th at co nfr o nts us no w is \,·hether possession among the
Greeks displ ayed certain cha racteristics chat w ould explain why Plato ig-
nored what I see as one of its uni versal characteristics: identification with
the god. If chis w ere the case, then the Greek system would be an excep-
j' tio n chat would certainl y ha ve to be taken into account. If chis were not
i the case, then we w ould first hav e co pro ve it. This is w hat I will tr y co do
I
n O \ V.

Dancing and Identif ying Oneself with the Gods

Both the vocabular y and the facts demonstrate that telestic m ania as de-
scribed by Plaro is indeed possession trance in which the subject identifies
himself with the god possessing him.
The verb ko rybanti ao, 28 translated by Baill y as, "co be agitated by a
Cor ybantic transport," according co Jeanmaire properly means "to Cory-
Music and Tr anc e amo ng the Greeks 207
Plato 's the ory The o ry in present work

ritual context

music and dance

and reestablishing and pro vok ing

movement by mo ve ment by
harmony with identification
itself itself
the Unive rse wi th the god

reinsertion in reinserti on in
the cosmos SOC!ety

healing

Figure 5 .1

bantize" or "to ac e like a Corybant," since korybant,ontes should be u ans-

laced, he says, as "Corybantizers" or "t hose who act like Corybantes."
An d wha t does "ac ting like Corybantes" me an if not beh av ing like the m ,
in o cher wo rds app arentl y identif ying wi th them internall y ?
Simil arly, the word baccheion, "Bacchant, " and the verb baccheuein, ' 9
wh ich Jeanmaire translates sometimes by "a ctin g like a Bacch ant " (19 51,
58 ), so metime s, mo re logically ic seems, by "act ing like Bacchus" (19 49,
466 ), express the ide a that by "ac ting like Bacchu s," in othe r wo rds by
"be ing a Bacchant," one identifies oneself with Bacchus . The process and
che state of identific ati on wi th the god possessing the entr anced person,
w hi ch are at the very center of any posses sion cult , are clearly expressed
by the Greek voca bul ary. But there is more.
Among the numerous epithets of Dion ysus , we find Bromius, the
"Roarer, " for he was the "bull-god, lion-god, earthquake -god" (D odds
1960, 7 4 ). In the Bacchae, describing the m ad running and danc ing of the
maenads in the mou nt ains (the oreibasia ), led by Dion ysus Bromius, Eu-
ripides h as the chorus say: "Whosoever leads the worshiping companies
[thiasos] is Bromius. " 3 0 According to Dodds (ibid. , 82 ), this mean s th at
the one who leads the oreibasia "is iden tified with the god. " And Dodds
adds : "Some d egree of identificacion wi ch the go d seems to be implied in
208 CHAPTER FIVE

ca lling human participants in Dionysiac rites Bacchae and Bacchoi ... it

is possible ... that the supreme degree of such ident ification, open only to .
the exarchos, was expressed by giving him the hol y name Bromius"
(ibid. ). Let us rec all here th at the exarchos was also the one who led the
dithyramb (Jea nm a ire 1951, 235 ). 31
Originall y, says D odds (1961, 83 ), rhe oreibasia" was mosr probabl y
·'a wom en' s rite w irh a single m ale celebrant," who was the exarch, an d
who as we have jusr seen identified himself wirh Dion ysus during his
trance. But wha t about the m ae nads? or hing wo uld lead us to believe
that they identified themselves wit h Dio nysus. Bur, as Gregoire (196 1,
225 ) obse rves: "Jusr as the God is a bull , hi s she-sectarists see themsel ve s
changed into cows, and we know th at th e Proitides 33 let out cow moos."
Why? Nonnos, in rhe D ion yszacs (8 )'' rep o rts rh at Semele, impre gnated
by Zeus and carrying Di o nysus, used to m oo whe never she heard th e low -
ing of a bull. Now rhen, as Jeanm aire not es (1951, 347 ), Semele is pre-
sented by onnos as being "t he model of the Baccha ntes. " We ma y thus
ju stifiably think th at whe n they were entranced, the maenads, or some of
them, identified rhemselves either with Semele, mother of Di onysus , or
wirh a cow. For hi s part, in h is introduction to his edition of Bacchae,
Dodds wr ites: "There mu st have been a time wh en the m ae nads or th yiads
or Bacchae really bec ame for a few hours o r days w hat their name im -
plies - wild wo men w hose human personalit y has bee n temporaril y re-
placed by another" (1960, xiv ). We could say the same of possessions in
Greece, which, at least o rigin ally, had no direct relation w irh Dion ysus,
and stemmed from pre-Dionysiac cults, th ar of rhe nymphs in particul ar.
All of this (together w ith the facts rhat w ill be cited subseque ntl y) re adil y
demonstrates, I trust, rh at rele stic mania, "inspired by Dion ysus" and
about w hich Plato speaks, is indeed a possession trance corresponding to
the definition given at the beginning of rhis book; in or her wo rds , it is
"identific ato ry."
Befo re ending on this point, we must add thar although relesric rr ance -
,,·herhe r iri s Dionysiac in rhe gene ral sense of the te rm o r pre -Di onysiac is
of lirtle m atter - is indeed "identific ato ry," the origi n alit y, the ne w ness,
and at the same rime the scan d alous aspect of trance tied to the cult of
Di o m ·sus srems from irs clo se associ ation w ith ,, ·ine and drunkenness .
Pentheus himself, the tr agic hero of rhe Bacchae, excla im s w ith virtu o us
indignation : "\'(!hen in festivities wine is ser ved to women, I say that then
there is nothing wholeso 1'!1ein their mysteries " ( l . 261 - 62 ). Symbols of te -
lestic t rance in Plato's days , the Bacchantes must also be seen as th e pr o -
moters of what we should call alco h ol ic trance, w hich in our d ays cer-
tainly is one of rhe most w idesprea d forms of mania. 3 5
Let us now turn to dance. Be it in w ritin gs or paintings, the Bac ch antes
and the m aenad s are usually repre sented as danc ing. 3 6 Evidence from vase
Music and Tr ance among the Greeks 209
3
paintings ; show that among the Greeks and elsewhere (and we have al-
ready spoken of this "elsewhere" in a previous ch apter ), possession dance
had two main aspects, whic h I shall here aga in call, abstract and figura-
tive. The abstrac t aspect is constituted by a repertoire of movements de-
void of any figurative function, in ot her words representing nothing other
than themsel ves: wild spinning, wide swinging of the arms , abru pt fling-
ing back of the head - the movement "that dislocates the neck" 3 8 described
by Pindar in a dith yramb - violent arch ing of the body, absent and convul-
sive attitudes : these are the principal - and well -known-fe atures of the
abstract part of possession dancing among the Greeks. As we can see (and
this is an often -repeated point, by Jeanmaire in particular ), there is great
similarit y here with the behavior still observed in so many possession
dances, from Bali ro Brazil, passing through the Middle and Near East,
the Mediterranean basin, and both black and whit e Africa. In this way,
savager y and aggression are externalized by means of scant gesticulation
that makes use above all of movements express ing agita tion and frenzy.
Presumabl y among the Greeks as elsewhere this was the way of expressing
the "fur y" of the god, and hence his terrible and redoubtable nature -
ulrimatel y, his po w er. Athenaeus, the Egyptian-born Greek who lived in
Rome during the later years of the second centur y A.D., whose Banquet of
the Deip n osop hists is an invaluable reposiror y of information on dance in
antiquit y, classified possession dances under the generic heading manio -
de is o rche seis, w hich we might translate by "frenzied dances." L. Sechan,
the gre at specialist of Greek dance, translates it as "da nses f urieuses"
(193 0, 83 ), making it quire plain th at he considered violence as its most
import ant feature.
Howe ver, possession dance also had another equall y important aspec t,
w hich consisted in executing figures that were either symbolic or imita -
tions of the beha vior of certain divinities or daimon es, and also of certa in
an im als . For Plaro, ro imitate "Nymphs, Pans, Sileni and Sat yrs" by means
o f mime is the ver y definition of"B acchic d ance " (Law s 815c ). We have the
evidence of a famous cup, made at the rime of the Median wars , that de-
picts a figure with the dress, hair, and accoutrements of Dion ysus -Bacchu -
s, pla ying the lyre, dancing \\·irh his head flung back in the characteristic
trance attitude, and flanked by two naked and staggering Bacchantes. Is it
the god himself who is thus represented) Is i~a Bacchant , and thu s an ex a r-
ch on? 39In this last case , it would mean that Dionysus also could be imitat -
ed while dancing. But Dion ysus was not alwa ys represented in his human
form. There is evidence that in Thrace, if not in Athens, during certain
ceremonies people wore ox horns in order ro imitate the Bull-Dion ysus
(Pearson 1918 , 127 ). The panther, anorher wild animal, was likewise imi-
tated by means of dance. There is a red-figure vase that depicts a maenad
dancing with a pair of clappers in each hand and wearing a panther sk in .
210 CHAPTER FIV:

"Is it not clear, " Sech an observes of this figure, "tha t she resembles chi
great wild animal less as a result of this external attribut e than on accoun
of the lithe and feline mo vement of her dance, w hich must have been pu~c
mated, imagination tells us, w ith sudden catlike bounds" (1930, 74 an ,
fig. 11 ). One cannot help but think that some maenads were possessed b
a panther at one moment or another of the ceremon y.
It might seem surprising that a maen ad-a companion of Di onysus, on
of those he addresses in the Baccha e (55 - 56 ) with the words "O my sister
hood of wo rshipers, whom I lead with me from barbaric countries, ..
w ho live and travel at my side" - can be possessed by a panther, or rathe1
for this seems to express better the spirit of the system, by the soul of
panther. Without mak ing an y diffusionistic connection between the tw
facts, I wo uld like to cite an exactl y parallel case obser ved in Dahome~
w hen, during a ceremony for Sakpat a the divinity of earth and smallpox,
woman embodying one of the "kings of the waters" (t9 x9s u ) before m
very eyes publicl y performed a panther dance, thus demonstrating that "
th at moment she was possessed by thi s wild animal , and consequent!
identified herself w ith it. The fact that in Greece a wo man embodying
companion of Dion ysus is depicted as possessed by a panther, or th at i
Daho me y a wate r divinit y fleetingly em bodies the spirit of a wild anima
sho ws, once again, th at the themes of po ssession are uni versal and that th
fury of w ild bea sts is one of its favorites. It further demonstrates th at po i
session is accomp anied everywhere by a broad mo vement of exch ange, b
a coming and going of souls, ' 0 a process that is certainl y within the logic c
a system whos e mainspring is a change of personality.
Th e dance repert o ire of the maena ds has ano ther aspect as well. The la
ter are depicted per forming what Maurice Emm anuel ha s called t!-
"joined hands dance " because of thi s ch aracteristic gesture that still re
mains enigmatic (unless I am mistaken ) bu t that und oub tedl y had a preci '.
symbolic meaning . As for the dances performed by those w ho imitated s;
tyrs and silen i, they we re made up of a variet y of steps. The se dancers, a,
cording to Sechan, "also on occasion practiced the dance w ith a forwar
bendin g and b ack archi ng of the bod y. But in general they favored abru1
crouching on one leg, followed by leap s that straightened the bent leg an
bent the pr eviously str aight one, wild cape rings in w hich the y often he
maen ads as partners, together w ith va rious contorsions and sway ir
movement of the hip s accentu ated by angul ar movements of the arm:
(ibid., 173 ).
Here , howeve r, a cau tionar y remark is in o rder. Jeanmaire, whose a1
thorit y on the subject of Dion ysianism is w ell known , asks (1951, 292 )
in their analys es of these orgiastic dances, the authors I have been quotir
(and others too ) ha ve not perh aps comm itted the erro r of "sometimes re
soning as if the choreography of the Bacchants and Bacchames was
t,
Music and T ranee amo ng the Greeks 211
learned cho reography, or at least one th at was r egularly ta ught and w hose
elements could be broken down into a set series of steps and atti t ude s." In
other words, for Jeanmaire the Bacchic dance could only be di so rde r. In
fact, besides the evidence already cited, ocher arguments also run counte r
to his thesis. T he reader will pardon me, I hope, for taking him once again
to Benin, but as an eyewitness I can confidently say that in this cou nt ry the
possession da nce performed during public ceremonies - wh ich often con -
stitute what amounts to theatrical perfo rmances: opera-ballets whos e
sumptuousness is displa yed in a large square specially set up for the event,
and before an audience of several hundred peopl 1c- makes use of an ex -
tremel y complex choreography whose steps an d figures, diligently
learned, are all charged with symbolic meanings . This applies, moreove r ,
not o nl y co the orisha cult (Yoruba ) bur equally to the vodun cults (Fon
and Gun ), and also, in fact, co man y ocher African cults, such as that of
the holey among the Songhay of. iger, for example, as Jean Rouch will
attest, or that of the zZir as described by Riya Salim a in H arems et M usu/-
manes d 'Eg ypte . '' We have no reason co believe that among the Greeks,
where possession also displa yed man y institutionalized aspects, the D ion -
ysiac dance was any less "learned" than ocher dances , whether religious or
profan ·~, or rhat it consisted of no more than a few elementary movements
convev'ing agi tation and frenzy. The indications I have listed so far tend tO
prove just the contrary, and comparison with what we know happens else -
where demonstrates that this was simply not the case. On this point, Jean -
maire seems to have been less perspicacious than he usually was. ow
that I ha ve concluded this digression - and it was nor without impor -
tance - I shall return co my argument .
In the an of dance, "orchesrics," the Greeks made a distinction between
rwo fundamental elements (Sechan 1930 , 64 ): the "movements" on the
o ne hand (phorai ), the "gestures, figures, attitudes" (schemata ) on the
other . We may reasonably presume rhat within the context of possession,
phorai served co denote the nonfigurarive movements (the abstract aspect
of dance ) since it is clear rhar schemata or " figures" denoted its figurative
aspect; Placo uses the latter word in the Ion (536b ) when describing the
dance of the "corybantizers" who, in order co conform co the tune of the
god possessing them , "have no difficu lty finding the right gestures [sche -
maton] and speech." We have already seen how va ried these schemata
were. Contrary co what Placo would have us believe with his usual ill will
coward relestic mania, the y did not consist only in "miming the actions of
drunk people ." Of course, one would expect that in Dion ysiac rituals
drunkenness would indeed be a habitual theme of imitation, the mimesis,
which was an essential feature of dance among rhe Greeks. Plato tells us so
himself (Laws 816a )." 2 Bur Bacchic orchestics was by no means confined
co drunkenness and frenz y. If it had been, why would Placo have stressed,
212 CHAPTER FIVE

as he does in the Ion (536b ) that "co rybantizers" (read: the possessed in
general ) "have a sharp ear for one tune only," the one w hich belongs to the
god by whom the y are possessed. And what merit would there be in their
"respondi ng freely" to that tune w ith gesture and speech, unless those fig-
ures were both va ried and characteristic? Here Plaro is flagrantl y contra -
dicting himself.
Indeed, it seems that Plato's ambi va lent feelings concerning telestic ma -
nia - or if one prefers, possession trance - lead him to contradict himself .
On the one hand he celebrates mania in general as being a "gift of the
gods"; on the other, he takes every opportunity ro cast a negative light on
the behavior of telestic mania . To be sure, he considers the latter as being
both sickness and the way to cure sickness, but he does not attribute the
curing process to the intervention of the gods, who are only asked to be
"benevole nt. " Rather, he attr ibutes it only to the dialectics of mo vement,
which is a purely profane explanation. Ultimately, the real p ara dox in Plato
is that he withdraws the gods from possession, or more precisely, from its
therapeutic dimension, whose importance he himself fully recognizes.
Th e reason is that, as Dodds has shown (1951, 217 - 18 ), he is di vided be-
tween his search for rational ity and his desire ro find a slot for the irra -
tional in his system. His moral fiber disapproves of the excesses, frenzies,
and unbridled behavior of telestic trance. Everything happens as if wi thin
him self, Plato does not reall y believe in possession; as if he cannot reall y
admit that these "mad" behaviors are uue to the presence of a god-or
rather a daimon, since for him the only real gods are the Demiurge and the
planets.
Despite its contradictions, Plaro's theor y does have the immense merit
of showing that possession essentially is a process through which the indi-
vidual is reinserted inro the whole that surrounds him, and, as a corollary,
that the role of music and dance is to reconcile the rorn person with him -
self. Whether the "whole" referred to in this reinsertion process is the cos -
mos , as Plato thinks, or the societ y, as I do, is of seco ndar y importance .
Whether in the first case reinsertion is due to harmon y with the universe,
resulting from movement, or whether, in the second case, it is due to iden -
tification \,·ith a god, the primary function of dance, is also of little impor -
tance . In both cases, the "symbolic efficacv, " as Levi-Str auss would say, is
of the same order and d irection.

To end, let us note here, since this in volves another aspec t of the mime -
sis of which Plato spoke, that imitation of the gods occupies an essential
place in Strabo's "theological excursus" (book 10 of his Geography ),
which acco rdin g to Jeanrnaire ( 193 9, 596 ) constitutes a veritable "theory
of orgiast ic cults ." A product of "ent husia sm," "the state of happiness indi-
cated by the wo rd eudaimonia , is, a bo ve all, imit ation of the divine state,"
Music and Trance among the Greeks 213
Jeanmaire wr ite s in the summary he gives of thi s text. Defined as "the art
of the Muses in the broadest sense," in other words including "dance, mu -
sic, and song," music is regarded by Strabo as a way of establishi ng "co n-
tact w ith the divine by means of the pleasure it procures and because it is
arr at its most perfect."

Plato and Possession Music; the Aulos

Curious though it may seem, the word music (mousike ) does not occur
once in Plato's writings in the context of possession. The only word ap -
proaching music - and approach ing it very closel y, but differing neverthe-
less- is mousa, which occurs once , and only once, in the Laws (790e ),
and wh ich I translated earlier as "musi a," for reasons I gave at that rime.
Plato speaks too willi ngl y of music and uses the word itself too frequently
elsewhere for this fact not ro be significant.
Whe n referring to what I am calling music in this book - and how,
among the Greeks or anywhere else, is one ro ralk about possession with-
out mentioning music? - what terms does Plato use? As we saw, he uses
"musia," "a ulos" (in its va rious forms ), "melody," and "rhythm." Of these,
however, aulos is by far the most frequently used. In direct or indirect" ' as-
sociation with trance , it occurs ele\(en rimes in his works, whereas melos,
in this or a derived form, appears only four times'" and the word rythmos
once.' 5
It is import ant to stress that whereas Plato links trance closel y with
melody or a melodic instrument, he rarel y assoc iates it with rhythm . In
connection with trance, nor once does he make any reference ro such pe r-
cussion instrument as drums, cymbals, or clappers, all of which are fre-
quently mentioned or depicted in this context elsewhere. Again, this fact
is significant .
For Plato, it is the rune, the melod y played by the aulos char counts. It is
when he hears it that the "co rybantizer" begins ro dance . Plato makes no
mention either of the sound intensity of the drums or clappers (instru-
ments rhe Greeks must surely have played just as loudl y as other peoples )
o r of the frenzied aspect of the rhythm (which also must have gone
through the same parox ysmic phases as elsewhere in the wo rld ) as the
dancing figures painted on Greek vases eloquently arrest. Indeed , Plato ac-
cords such slight importance to rhythm in possession rituals that w hen it
comes to identif ying rhythms, or more precisely, to saying which of them
"is appropria te to illiber ality and in solence or madness or other evils" (Re-
public 40 0 b )- since that is how these things appea r to him, we must re -
member - he simpl y refers the reader to Damon and passes on . In fact gen -
erally speaking, the ethos, as it was to be called later, of rhythms is of
scant interest to him altogethe r: "But which [rhythms] a re imitations of
214 CHAPTER FIVE

which sort of life, I am unable to say" (Rep. 400a-b ). When Plato writes
about possession , it is ro the aulos first and foremost that he makes refer-
ence. Rhythm leaves him indifferent, and he says nothing about percus -·
sion instruments. Arisrotle was ro rake much the same course in the Poli-
tics (8 ), in which he deals with rhythm merely in passing, and as a wholly
secondar y aspect of music seen in the context of its relation to enthusiasm.
Bur it is not onl y in the work of these rwo great theoreticians of the rela-
tions between music and trance that the aul os is presented as the instru-
ment par excellence of mania . The same is true for the works of the tragic
poets. Aesch ylus, in the onl y survi ving fragment of his traged y The Edon-
ians, "6 mentions the bomb yx and of its melod y (m elos ) whose call "leads
to madness [m ania]." (The bomb yx was a type of aulos, a single clarinet
with a long bod y and remo vable reed; it had movable metal rings - bom-
by ko i- that acted on the side holes rather like modern keys. 47 Sophocles,
for his part, says of Ajax, when he was seized by a murderous frenzy, that
he was liter ally "synaulized " by "div ine m adness" (Aja x , 610 ). Euripides,
in the Bacchae, specific ally names the aulos several times, and also the lo -
tos (160 ), which was probabl y a particul ar variet y of the aulos. Burro be
honest , although in thi s tr aged y wholl y centered on possession he talks of
the aul os as an instrument typical of Dion ysus worship , he also sa ys the
same o f the drum , "th at h ide -stretched orb" (ibid ., 123 ), and nothing is
ever explicitl y said about the exact relation that either instrument had
with trance. In H eracles, on the other hand, the murderous frenzy of the
hero is directl y related, three times in succession," 8 co the aulos. Moreover,
it is explicitly stated that no drums will take part in the sinister dance that
Lyssa promises (H eracles, 891 ). Thus the bloodthirst y mania that grips
both Ajax and Heracles, when the latter are in the power of Lyssa, the
"r age ,'" 9 is pre sented as linked w ith melod y and not, in an y wa y, with per-
cu ssion.
All this, o nce again , contr adicts the myth that trance is triggered b1
drumming and its supposed effect s on the br ain 's alpha rh ythm. The
trances with which w e are dealing here ha ve a quite different explan ation .
despite Neher 's and his disciples' opinions. ' 0 The unanimity shown by Plate
and Arisrotle on the one hand , and by Aesch ylus , Sophocles, and Euripi ·
de s on the other, in presenting the aulos as the instrument of trance is cer
t ainly signific ant, and it must reflect what was commo nly belie ved at thi
time a nd ultimatel y, the actual musical practice s current in Greek posses
sio n cult s . W hether a n aulo s, bombyx, o r Iotas , it wa s a melodic instru
ment that the Greek s credited w ith triggering trance. Neither rh ythm, no
she er volume , nor the fren zy o f the percussion instruments - all of whicl
w ere frequentl y used in posses io n ritual as we know from so man y paint
ings- were seen as ha ving this effect.
W h at exa ctl y w as the place of the aulos a mong the ancient Greeks
Both textu a l a nd pictori a l evidence clearly si10w th at it w as closel y assoc i
Music and Trance among the Greeks 215
aced with trance, but that it was not linked to it exclusively, and tha t it was
not reserved solely for possession rituals; far from it. In fact the aulos was
a ver y popular instrument in Greece, at least in its simple forms; we may
well suppose that the instruments of more complicated construction, suc h
as the bombyx Aesch ylus mentions , were pla yed onl y by mp -flight profes-
sion als . And although after the fluctu ations in popul arit y it underwent in
Greece at the time of the Persian wars-Aristotle (Politics 8. 6 ) explains
this ver y cle arl y- it did become the object of a kind of craze among the
upper clas ses,5' it was a generall y despised instrument. Whether the y were
men or women - for it is often depicted as being pla yed by women -t hose
w ho pl ayed it during banquets were either slaves or people of very low es-
t ate. What Eri x imachus, in the Sym pos ium (176e ) says w hen he proposes
"that we should send awa y the flute girl [aulos girl ] who has just come in-
let her play to herself or, if she likes, to the women of the household,"
clearl y demonstrates the small regard accorded to aul os pla yers. This
same auto s pla yer w as to return later (Symposium 212 ), accompan ying a
band of re veler s cre ating a gre at din and supporting a ver y drunk Alcibi -
ade s. The role assigned to her is significant. In about 5 00 B.C. , in other
w ords more than a centur y before the Symp osium w as written , Pratinas of
Phlius, the master of the Athenian sat yric chorus, had alre ady wri tt en:
"Son g , the Muse had decreed it king, let the autos keep its second rank for
it is its ser vant. Let it onl y take command in unrul y processions , and the
fistfights in which drunken young men indulge on the doorsteps of whore -
ho uses. " The aulos was thus , broadly speaking , linked with debauchery."
It was also associated, nonetheless, with militar y exercises, with war,
and consequentl y with courage. It accompanied the pyrrhic dance, which
had spread from Crete, said to be its countr y of origin , ro all of Greece,
and in p articular to Sparta, where it w as con sidered "as a propaedeutics of
w ar " (M o utsopoulos 1959, 146 ), and to Athens, w here "it enjo yed great
succe ss in the Panarhenian festi vals." Plato w as part icul a rly appreciati ve
of thi s d ance w hich was responsible for "ph ysic al beaut y and nobilit y of
cha racter in th e young, bec ause it imit ated the gesture s o f va liant war -
rio rs" (ibid. ). W hen acco mp anying the p yrrhic d a nce , the .wlos, it is true ,
co m bined it s sound with th at of the lyre , but it w as nevertheless so closel y
identifi ed w ith thi s d a nce th at "it wa s re ferred ro indifferentl y by the two
term s aulos and pyrrhikine" (ibid .). Moreover, a vas e depicts an aulos
p lay er in the mid st o f armed , charg ing wa rriors, pl ay ing his pipes to w ard
t he sky (Sech an 193 0, 107 ). Lastl y, the au tos w as also so met imes associ-
ate d w ith a sow ing d ance (ibid. , 100 ).
The au los appears, then, to ha ve been a p opular instrument associated
not onl y with trance but also with the the ater, debaucher y, wa r, and agrar -
ian rites. The least we can say, therefore, is that its use w as not a special -
ized one. As ro its timbre and the m anner in which it wa s pla yed, we can
sa y, without fear of error, that the y were as piercing and as shrill as they
216 CHAPTER FIVE

still are throughout the Mediterranean basin in our own day. Paintings de-
pict aulos pla yers blow ing their instruments with puffed-out cheeks wh ich
indicates that, as is almost universal with this instrument, the y practiced
circular breathing, which enables one ro play without taking breaths, and
thus w ithout sropping. As played in the Mediterranean manner, let us say,
the double clarinet and the oboe have very vf'.hement intonations, a force -
ful and rasping rone, and an emotiona l intensity enhanced still further b)
the fact that the instrument can be pla yed for hours on end without inter-
ruption. In Turke y, a bagpipe - another form of the clarinet or oboe-ii
used to incite camels ro batt le . It is tempting to think , in view of its shril
vol ubilit y, its eloquence, its intensit y, wh ich makes the air positively vi
brate and can be deafening, and its insistence, w hich can become almos
obsessive, that this instrument is particularly suited to the cask of makini
people's he ads spin and to triggering trance . This would explain the clos
association of the aulos w ith possession. In fact, such an interpretat io1
might as well be the pure product of our imagination. othing proves th a
the sou nd of the aulos is experienced in this way by people actually he arin
it, and nothing in the texts demonstrates that this was the case in ancier
Greece . To be sure, Aeschylus, as we just saw, says that its call s "lea d t
madne ss," but Euripides in the Bacchae (127 - 28 ) speaks, on the contrar
of "the sweet calling breath of Phrygian flutes [read aulos]." 53 If the role c
the aulos in triggerin g trance had been due to its musical characteristic
and in particular to its timbre, then its rel ation to trance would have bee
doubl y exclusi ve . Its use would have been reser ved solel y for possessi c
and, reciprocally, po ssessio n wo uld have been induced by no other mean
But this was nor the case. We ha ve seen th at the aulos was used in a gre
many other contexts. T ranee likewise was occasionally accompanied by
quite different instrument. We know th at in Greece the opposition b
tween aulos and zither, betw een aulos music and zither music, berwe ,
Dion ysus and Apollo was strongly emphasized . And yet, as surprising a~
m ay be, we find that rhe lyre could occasionally be associated with Bacd
tr ance . Thi s is shown by a paint ing I mentioned earlier, the one on the re
figure cup in the style of Brygos and dating from the rime of the Persi
wa rs ,'' repre sent ing Dionysus in a sta te of trance and pla ying a seve
stringed lyre. Once again we see that in ancient Greece as elsewhere, t
relations between music - here the instrument of mu sic- and trance W<
not ruled by nat ur al constraints but by cultural arbitrariness. It is nonetl
less true that thi s representation of Dion ysus hold ing a lyre is surprisir
Why is the god playing this instrument? To what myt hological cont,
does this refer? For all I know, the question awa its an answer.
Let us now pass on to Plato's thoughts on the aulos . Tak ing as his st,
ing point the passage in the Sy mposium (215c ) in which we are rold t
the melodies 5 5 of Marsyas played on the aulos are the only ones capabl
Music and T ranee among the Greeks 217
inducing possession by reason of their divine origin, Moutsopoulos
( 19 59, 10 ) conch 1des th at "thanks to chis reli gious aspect of their art, au -
los pla yers, having been banished from Th e Republic, reappea red in the
Laws" - a conclusion somewhat swiftl y drawn. It Vl·ould be very surpris-
ing if Plato, whose feelings concerning trance were so mixed, had seen in
its association w ith the autos a good reason for rehabilit ating the instru -
ment. The matter should be viewed somewhat differently . Th e autos is in-
deed condemned in Th e Republi c (39 9c- d ), bur this judgment is only
made as part of a general condemnation of all "pol yharmo nic insrru -
ment s , "56 a condemnation char applie s just as much to "instruments of
man y strings" as to the autos- the most 'polycho rd' of all." And Plato
asks: "Do not p anharmonic instruments imitate the autos?" We know char
Plato accused autos players of ha ving introduced all sorts of innovations
into music, notabl y new "harmonies " requiring more notes than the only
rwo m oJes char "should be kept," char is to say, the Dorian and the Phr y-
gian. In chis passage, then, Plato has in mind instruments of recent and
complicated workmanship, furnished with a great many side holes (possi-
bly the bo mb yx with its long rube and mobile rin gs mentioned earlier ).
Bue he is no t attacking the simple autos, with its small number of holes,
any more than he is the lyre or the zither when they have a small number of
strings. He is attacking the virtuosos and their sophisticated instruments
chat upset tradit ion. He is not attacking tradi ti onal in struments in any
w ay. And chis explains wh y, when he is dealing with the organization of
competitions "for both men and horses" (Law s, 764d - e), he groups
"rh ap so di ses, zither and autos pla yers" together and chink there should be
"a separate umpire " for chis group . le is therefore nor quire correct to say
rh a,c the autos is banished fr o m Th e Republic and reappears in the Laws. If
w e look more closel y, we see char a different instrument is involved in each
o f the t w o cases : in one it is indisputabl y the instrument pl ay ed by virtuo -
so s; in the other it is the tr adition al folk instrument. The two were ver y
different.
1onetheless , the musical instrument dear to Plato w as the lyre or zither ,
in sho re a stringed instrument (as long as there were not too man y
strings! ). By their very nature as stringed instruments the y were associated
with ideas concerning the simple relations of length s me asured on a vibrat -
ing sering, and hence to mathematical speculation about numbers; thereb y
the y were connected with the highest aspect of music, the one concern-
ing the relations of man with the cosmos and the great laws of the uni -
verse.
The au tos, on the other hand, is a music ally approximate instrument,
and the relation between its sounds - sounds that are easil y al ter ed by
blowing differentl y or by differentl y co vering the side holes - are more dif -
ficult to express in numbers. In the Phitebus (56a ) we read char the arr of
 218 CHAPTER Fl'

the aulos is one of adjusting its harmonies not by measurement but l

"empirical conjecture." Added to its association with trance and debaucl
ery, this aspect of the aulos would certainl y not have led Plato to rank
high among musical instruments. Ho weve r, in his respect for traditic
and traditional music, he accepts the aulos wi lly-nilly, and carefull y marl
its place in the Laws, as we just saw. Meanwhile, his real feelings bur
forth in The Republic (398 - 99 ) when he proscribes it, without stipulati1
that he is in fact referring to the modern, virtuoso instrument. Forgetti1
what he has just said about "instruments of many strings or whose cor
pass includes all harmonies," and which "we sha ll not need in our son
and airs" because the y are responsibl e for "dirgelike modes of music th
are useless even to women," he asserts that "in preferring Apollo and t:
instruments of Apollo to Marsyas and his instruments"-meaning the ly
and zither to the various forms of aulos - h~ is simply "purging" the city
the "luxury" infecting it.-
One could not judge the aulos more severely. Plato, nevertheless, m
cured a render feeling for this instrument. Witness in the Crito (54b ) t
last words he has Socrates utter before his death . Resisting the appeals
Crito, who urges him to flee, Socrates decides to submit to the sentence
his judges, however unjust it might be; he tells Crito what the voice of I
conscience is dictating to him and concludes: "Th at, my dear Crito,
what I seem to hear, just as it is said that those who corybantize seem
hear the aulos." As we see, Plato 's feelings about the autos were as mix
and as contradictory as those he had for possession.
As for Aristotle, he was not very fond of the aulos either, though for c
ferent reasons. Recalling the story (Politics 1341 ) of Athena who thr ,
away rhe aulos after she invented ir because, as she pla yed it, she saw t
reflection of her disfigured face in a spring (wh ich showed that she ~
blowing it with puffed-out cheeks, the usual technique of aulos pla yin
Ar istotle notes that although the goddess certainl y did this out of annc
ance becausy of the ugl y distortion of her face, the real reason was ti
:~t "education in flute playing has no effect on intelligence," something ti
, :1
I; would indeed disgust Athena, the inventor of the sciences and arts! (1
r,: I ',. cording to another version of the legend ,5' Athen a flung the pipes a~
' from her with these words: "Far from me, shame and taint of my bod
do not indulge in such indecency!" whic h again is hardl y compliment ;
to the aulos. )
Elsewhere, again in the Politics (1341a ), other reasons lead Aristotle
depreciate the aulos. This time, the instrument - which should be w
only when "cathartic rather than educational effects" are requirec
should be excluded from the educational training of young men beca 1
one cannot speak while one is pla ying it; this ob viously implies that sue

.......
Music and Trance among the Greeks 219
defect does nor apply to the lyre or the zither, since both are the inst ru-
ments par excellence for accompan ying singing. The aulos is thus co n-
demned to be an instrument pla yed onl y by professionals, slave mus i-
cians, or people of low estate. In fact, Aristotle's entire chapter on the
educ ation of young men, and particularl y education through music (in the
Politics ), is impregnated through and through with a very marked class
prejudice. Nobles and common people , sla ves and free men are constantly
being contrasted: "But professional musicians we speak of as vulgar peo-
ple, and indeed we think it not manly to perform music except when
drunk o r for fun" (1339b ), and furth er on he writes: "we do not consider
performing to be proper for free men " (1341 b ). Still further on (1 342a ),
dealing w ith theatrical music, he concedes the necessit y of pro viding "ca-
th artic " harmonies and melodies (meaning Ph yrgian, and thus suited to
the autos ) in order to satisfy the tastes of "the vulgar class composed of
artisans and laborers and other such persons."
Plato , as we ha ve seen, had little esteem for Bacchic dances-
accomp anied by the aulos- which are unsuitable for citi zens. " We also
saw the lack o f respect he had for the aul os pla yer in the Sym po sium. In
the T heaete tus (173d ) he associates the instrument w ith the hetaera. For
Plato as well as for Aristotle, both of high social status , the aulos was con -
temptible bec ause it w as the instrument of the lower clas ses and of sla ves.
In a sen se, the aulos had in ancient Gr eece the same position the accordion
no w has in Fr ance. As for Dion ysiac so ngs, the y were essentiall y popular,
as N ietz sche saw quire well in The Birth of Tr agedy . Here once again we
find w h at w as sa id earlier abo ut rhe highl y popul ar character o f posses -
sio n mu sic .
Let us return to the rel ation o f the aulos ro mrznia. Ap o llo wa s seen
among the Greeks as the inventor of rhe"lyre and M arsyas the Silenus as
rhe inventor of the aulos . And about rhe mel o die s" o f Mars yas, Plato
catego ric ally says : "w hether pla yed by a fine aulos pla yer o r a paltr y auto s
girl , the y are the o nl y o nes capable of inducing possessio n, bec ause of
th eir d ivine origin, and to 'indic ate th ose w ho are recipients of rhe deities
an d their [proper ] rites " (Sympos ium 215c ). Wh y is this' Because the y
"a re themsel ves di vine. " An explan ation as unexpected as it is peremp -
tor y ! Bur wh o pro vide s it? Alcibi ades , who is drunk and delivering his
celebr ated and highl y ironi c speech in praise of Socr ates. In other words, it
sh ould not be taken seriousl y. In short, Plato tells us th at tunes on the au-
tos en r anee pe ople, and this is corr obo rated by to o m any other sources
for us to doubt it; bur he abstains from explaining this particul ar effect of
aulos music. This fact is important and deserves to be stressed. Although
Plato has a theor y of the effects of music and dance on the curing of mad -
nes s, as we ha ve seen , he lacks one to explain the effects of music on the
220 CHAPTER FIVE

triggering of trance ( though as we shall see, others would later formulate

one in his place ). Arisrotle, on the 9ther hand, had very precise ideas on
the subject, so we will now turn co him. ·

Arisrotle, the Ethos of Modes, and the Phr ygian Mode

Plato tells us that if the "productions " of Mars yas ha ve the strange powe1
of causing trance, this is "be cause they are themselves di vine." Hi s expla-
nation is a bit brief. Arisrotle proposes a different view of the matter. In the
Politics (1340a ), ha ving talked about the tunes composed by Olympus,5 '
"which make our souls enthusiastic, " and ha ving given a long account 0 1
the imitative virtues of music and the nature of "harmonies" - or musica
modes - he makes the general statement that "the Phr ygian mode makei
men enthusiastic [enthusiast ik ous]" (1340b ). Marsyas the Silenus and hii
pupil, Ol ympus, are both, as we know, Phr ygians. The aulos is Phr ygiar
in origin. If the tunes of Mars ya s and Ol ympus induce trance , they mus ·
do so because they are in the Ph rygian mode - a more interesting explana
tion, it will be conceded, th an Plaro's . And all the more so because it is par
of a general theor y of music 's effects based on the idea that music is able tc
"represent" and "imitate" states of the soul. Melodies (Mele ) "do actua ll:
contain in themselves imitations of character" (mimemata ton ethon ), Ar
isrotle tells us, and in "melodies there are differences, so that people wher
hearing them are affected d ifferentl y and do not have the same feeling :
with respect co each of them" ( 1340ab ). The same is also true of rh ythms
he adds. Returning a little later on to the Phr ygian mode, he write
(1342b ): "the Phr yg ian mode ha s the same effect [dyna min] among har
monies [harmonion] as the aulos among instruments, bot h are orgiasti ,
and passion al [orgiastika kai pathe tica] . .. all Bacchic transport [bac
ch eia] and all movement of this sort belong s to the aulo s more than an :
other instrument , and find their suitable acco mpaniment in tunes in th ,
Phr ygian mode among the harmonies. "
Being "passi onal " or "patheti c," the Phr ygian mode is thus opposed t
the Dori an mode , whic h is "ethical." Properl y speaking, then, the Phq
gian mode is not endowed wi th ethos or moral character, but with pathc
or passion. For this reason, unlike melodies in the Dorian mode, whic
th e m an or youth of high birth m ay learn because the y are "eth ical" (eth
kotatai ) and there fo re suita ble for educational purposes, those in th
Phr ygia n mode mu st only be listened to, and their performance must I:
left to mu sicians of ser vile or low condition. It is worth noting , in passini
th at here w e return, th ough by a n unexpected p ath, to a fact we est al
lished in an earlier chapter, n amel y th at possession music is mu sic one Ii:
tens to and th at is pla yed by others.
Music and Tr ance among the Greeks 221
Aulos music is Phr ygian , then, and it excites "ellthusi asm" amo ng those
persons in whom the effect of rhe "sacred melodies" (hieron melon ) incite
rhar kind of emotion . Those persons are then thrown into a stare compa -
rab le to that produced by admini stering a remed y or purge (katharsos),
and the result of these "purgati ve" or "cathartic" melodies is to produce a
feeling of liberation in tho se to whom th ey are admini ste red. Briefly sum -
marized, thi s is the well -known theory of the re lations bet ween music and
trance th at Aristotle advances in book 8. 7 of the Polit ics . 6 ° Clearl y it is
very different from Plato's. Not a wo rd here of either d ance or movement:
everything centers upon a theor y of the effect (dynamis ) of music, wh ich is
ethical w hen the mode is Dorian , orgiastic when the mode is Phr ygian. In
Plato , thi s aspect of the matte r is nor even mentioned, the reaso n being
that he does not hold the same views as Aristotle on the etho s of modes. 61
Aristotle is at the same time clearer and more liberal. For him, rhe Dorian
alone is "ethic al, " which is to say moral and worthy of figuring in the edu -
cational program of well-born you ths. The Phr ygian , being "o rgiastic "
and "p ass ionate, " must be excluded from such a program , and should only
be listened to whe n pl aye d by employees of low estate or slaves . This said,
however, all modes are good provided the y are used in the correct way at
th e right time .
In Plato, thing s are rath er different: he too pl aces the Dori an ab ove all
the other m odes, but hi s opinion of the Phr ygian, which is much less clear-
cut, fluctu ate s over the yea rs. In the Lach es (188 d-e ), the Dori an is pre-
sented as rh e only m od e rh ar "is trul y Greek," while the Phr ygian is pre-
sented as being as little wo rth y of respect as the Lydian and certainly rated
lower than the Ionian. In Th e R epu b lic (399 ), which is to say much later
in his life, Pl ato reject s both the Ion ian and Lydian as being "so ft" and fit
onl y for drin king to, so rhar the onl y t\ 'J O with wh ich he is left are the Dori an
and Phrygian, of wh ich he sa ys: "Lea ve us these rwo modes - the enforced
and the vo luntar y- that will best imitate the utter ance of men failin g or
succeeding, the temper ate, the br ave- lea ve us th ese ." Here, then, Plato is
putting the D ori an and the Ph rygian on the same footing. It is also wo rth
not ing that whereas, as we ha ve seen, Aristotle identifies the aul os wi th
the Phr ygian mode, Plato, on the contrar y never once associates th e words
"Phrygian " and "aul os." Similarl y, Aristotle's repeated emphasis on the
Phr ygian 's connection w ith ent hu siasm is m atc hed by Plato's total silence
o n this subjec t: the latter never once assoc iates Phr ygian music and trance.
And ju st as rhe ex pl an ation of trance onset through the effect of the ethos
of modes is central in Aristotle, so again it is entirel y abs ent in Plato .
Plato's an d Aristotle's interpretati ons of the relations bet wee n music
and tr ance are thus quire different. Plato 's theor y of move ment in har -
mony with the cosmos carries on the Pythagorean tradition . By system -
222 CHAPTER FIVE

atizing the ideas inherited from the past concerning the et hos of modes,
and at the same time relati vizing the m , which is to say giv ing them a m uch
more ps ychological than moral content, Arisrntle branches off in a qu ite·
different direction. We are not concerned here w ith decidi ng wh ich theo ry
is the better. What is interesting is to observe the extent to wh ich t he views
of the two philo sop hers differ on the subject of mu sic's relatio n to tra n ce,
even though they are separated by only about fift y yea rs , and t he y both in-
terpret what must have been currently o bser va ble facts presenting much
the same aspect to each of them.
For Aristotle, then, it is the Phr ygian mode that triggers trance, by its
sui generis virtue. This being so, a question naturall y poses itself: What
were the specific music al characteristics of rhis mod e that enabled it to
produce such effects? That is the question I shall now try to answer.
First, it should be noted that the term Phrygian undoubtedly changed its
me aning over the centuries in ancient Greece, and that if it designated a
certain mode in the time of Plato and Aristotle, it certainly designated an -
other one in the time of Aristoxenes. Needless to say, our quest ion applies
onl y to the, let us say, primitive Phr ygian mode , the one to which our two
philosophers referred. Let us thus turn to two sources dealing with data
considerably earlier than Ar istoxenes: first , the comparative table pro -
p osed by F. Lasserre ( 19 5 4 , 40 ) of the six Greek modes as they would ha ve
been at the time of Lasos of Hermione, which is to say at the end of the
sixth century; second, that presented by J. Chailley (1960, 42 ) of Plato 's
six harmonies as transcribed by Aristides Quintilian in the second centur y
.-\.0_.,
If we restricted ourselves to these d ata, which represent the traditional
views on the Greek modes, the question we raised would remain an enigma.
Aristotle m akes the mo st clear -cut distinction between the Do r ian and the
Phr yg ian with respect to their psychological effects. One would the r efore
expect the two modes to be equally distinct with respect to their musical
structures. But this is not the case. If we are to believe the se two rabies, the
D or ian and the Phr ygian \\·ould then both have the same sequence of in -
tervals (tones, quarter-tones, thirds ), wi th the single difference that the fi-
nal interval (going up the scale ) would be a major third in the Dorian and a
full tone in the Phrygian . Thus as Chailley observes ( 195 6, 15 8 ), the Phr y-
gian would have differed from the Dori an "sole ly by its final upper note, a
re instead of a mi." Lasserre's table, on the other hand, establishes a differ-
ence of rel at ive pitch between the two modes, with the Dori an being 2
fourth lower than the Phr ygian. Quintili an's table shows no difference o J
this kind. In an y case, this difference is not significant with respect tc
music al expression, since it does not affect the relations between the note !
that co nstitute the mode. Clearly, then , the Dori an and the Phr ygian, ac-
cording to our two tables, were almost identical in structure. T heir ex -
Music and Trance among the Greeks 223
pressive possibilities would thus also have been much the same. Only a
very experienced ear could have made a distinction between them. Under
these conditions, it is difficult to see why the Phrygian should have been
the mode of madness, let us sa y, and the Dorian that of serenity. That two
modes so indistinguishable from each other musicall y could have been en-
dowed with two ethos as different, an d indeed as opposed to one another
as Aristotle described them, rem ains incomprehensible.
For Aristotle (Politics 4. 3 ), since it is to him that we must turn once
again, the difference between Dorian and Phrygian is comparable to that
between a north and a south wind. We are not, therefore, dealing with
mere nuances. In his eyes (or ears ), the difference between what was called
the Dorian and Phr ygian in his time must have been musically ear-
shattering. As we ha ve seen, the traditional interpretations of what the
Greek modes were offer no explanation at all of this difference, but an-
other solution exists . Indeed, the very data underlying the problem ha ve
been recently renewed by Samuel Baud-Bovy's research based on a first-
hand study of Greek peasant music as it can still be heard today, w hich in
itself is an extremely interesting and innovative procedure. Commenting
on the coexistence in Greece of t\VO different musical systems, one penta-
tonic and without semitones (anhem itonic ), the other diatonic with semi -
tones, Baud -Bov y ( 1978, 189 ) advances this hypothesis: "that such an op -
position already existed in the music of ancient Greece, illustrated by the
rivalr y of Apollo and Mars yas, with rhe aulos of Mars ya s the Microasian,
a diatonic instrument by nature, opposing itself to Apollo's lyre, whose
strings, however many in number, certainly did not produce anyt hing oth-
er than sounds belonging to the an hemitonic systems." According to
Baud-Bov y,6; the primitive Dorian must have been "a pentatonic anhemi -
tonic mode, whereas the Phr ygia n was a diatonic mode. Apollo and his
lyre triumphing over Marsyas and his aulos would then symbolize a vic-
tor y of the Ancients over the Moderns." There is every reason to believe
that this Phr yg ian mode with semitones was a heptatonic. The Dorian
versus Phrygian opposition would thus be the same as that distinguishing
between a pentatonic mode without semitones and a hepcaconic mode with
semitones. Musically, rhe difference is considerable. It ir comparable to
the one made today in Java between rhe slendro and pelog modes. And it
is much more marked, for instance, than the already very perceptible op-
position between major and minor mode in Western music. Under these
conditions, one can readil y understand why the Greeks so clearly, and in
so many respects , distinguished between the Dorian and Phrygian. The
difference was quite perceptible.
We now need to know why this Phr ygian "harmony," ha ving the clear-
cut musical characteristics just described, was in Aristotle's terms "pathetic,"
"o rgias tic," and "enthusiastic"; in other words, since chis is our concern
224 CHAPTER FIVE

here, suitable fo r inducin g trance. Musically, it is clear th at a mode able to

make tone / semitone contrasts must offe r much gre ater expressive possi-
bilities th an one th at lacks this possib ilit y. The aulos, a Phr ygian instru -
ment, is obv iousl y more expressive than the lyre, a Dorian instrument.
Th e Phr ygian aulos pla yers, as Lou is Lalo y 6 ' obse rves, mu st have been the
"true gips y [fiddlers ] of antiqu ity." Th e instrume nt 's technical possibili -
ties, rhe qualit y of its t imb re , and rhe wealth of its inflections, in conjunc -
tion with the exp ressive possibil iti es of rhe mode , certa inl y allowed play-
ers to produce melodies of great emotional imp act. The Phr ygian
undoubte dl y was - and here we concur with Jacques Chailley - as much a
style as a mode. To use Isobel Hender son's terms ( 1957, 382 ), the wo rd
harmonia did not only des ignate a cert ain scale bur also an "idiom ."
Should we conclude from this th at Phr ygian melodi es we re so cha rged
w ith expressi viry and emotional content that the y were capable of induc -
ing trance by virtue of this qualit y alone' This would mean accep tin g rhe
idea that music has a power sui g eneris, an ide a refuted at lengt h in earlier
chapters . But this, one may sa y, is ne vertheless Aristotle 's theor y! Not
quite, I would rep ly. True , he w rite s (Politi cs 1340a ) th at the runes of
Olympus make men 's souls "enthusiastic" and enthusiasm is an affl icti on
(path os ) of the soul. Tru e, he repeats elsewhere (1342a ) that under the in -
fluence of these "sac red melodies" certain persons are "possesse d " by a
fo rm of ag itati on known as enthusiasm. Bur he presents this effect of the
Phr ygian onl y as a p articular - and , witho ut doubt, extreme - case of a
m o re ge ner al actio n tending to produce emotion in th ose inclined to "pit y
or terr o r."
The well -known pass age in the Po litics ( 1342a ) in w hich Aristotle 's
views on th e rel ati on of music to tr ance are exp re ssed, is q uire sign ificant.
The katharsis tha t occu r s in possession rituals, as a result of "enthus iast ic"
harmonies and "sacre d melodies, " is presented as being of the same order
as rh ar which is at work in rhe theater. (In passing, let me point out that
this is probabl y the earliest rext associating p os session and theater in thi s
wa y.) More gene rally st ill , these "purg ative melodies ," w hich arouse "a
pleas ur able feeling of relief'" in emotional people , provide men w ith
"ha rml ess delight. " The Phr ygian mode thus covers a ver y broad musical
field, one that includes trance music, the ater music, the equi vale nt of va ri-
et y show music (I mean those song s th at mov e soul s subject to "pit y and
terror,'" just cited ), and last , dance music, when rhe latter is agitated or un -
bridled in character. M o reover, as it emerge s from book 8 of the Politics,
this Dor ian / Phr ygian oppos iti on su bsumes a whole ser ies of ot her opposi-
tions: calm vs. agitated, \·irile vs. effemin ate, worth y vs. unwo rth y, aristo -
cratic vs. pleb eian, beaut y vs . banalit y, educational vs . entertain in g. The
Phr ygian side of this opposi tion could be summed up in one wo rd: rele ase.
For Aristotle, then, Phrygian does not mean onl y "enth usias tic" in the reli-
Music and Tr ance among the Greeks 225
gious sense of the term , but rather "org iastic " in its general and rather late
sense. When Aristotle w rites (1 341 a) th at the aulos -and therefore the
Phr yg ian mode- is not "a moralizing but rather an orgiastic influence," so
that it should onl y be used for purpo ses of "pur ification rather than in-
struction ," he is certainl y not thinking only of Dion ysiac ceremonies. He is
also thinking of theatrical perform ances or banquet s; in short, all situa -
tions in w hich people seek rele ase. On the contrar y, w hen he spec ificall y
targets posses sion (Politics 1339b ) and the music that induces it, he does
not mere ly t alk about the Phr ygian, but is much more specific: th e tunes of
Ol ympus are the ones that make our so uls "enthusiastic." Naturally, this
must be set alongside the pass age I h ave often quoted from the Symposium
(215c ) in which w e are told that only the tunes of Mar syas are capable of
"inducing possession ." W hy th is insistence on the two n ames ? Ob viousl y
because, by using them , Aristotle and Plato are indic ating ipso f acto that it
is not just any Phr yg ian melody or ju st any autos rune th at triggers trance.
By thus specifying th at o nly the tune s of Ol ympu s and Ma rsyas are ca -
pable of indu cing tr anc e , both w riter s obviously have a p articu lar reper -
toi r e in mind. What might its character isti cs have been? Th e texts tell us
nothing . But the refe rence to two legend ary 6 ' if not mythical cha r acters
implies th at the repert oire in question was anc ient and tr ad ition al. Th is is
not all, however. Marsyas and Ol ympus, an d p art icu la rly the fir st, are
strongly branded , if one ma y say so, as Phr ygians. To mention Marsyas is
almost to name Phr ygia . And what w as Phr ygia if not precisely the home-
land of Dion ysus? 66 "From the fields of Lydia and Phr ygia, fertile in gold, I
traveled first to the sun -sm itten Per sia n plains, " he tells us himself in the
Bacchae (14-15 ). And a little later: "O my sisterh ood of worsh ipers, whom I
lead with me from barbaric count ries, from T molu s, bast ion of Lydia, wh o
live and tra vel at my side. Raise the music of your own countr y, the Ph ryg ian
drums invented by Rh ea the Great Mother, and by me." And lastly, in a later
passage the chorus describes th e Corybantes as "wedding th eir frenzies to the
gentle breath of the Phrygian pipes ." In the time of Plato and Aristotle, "Phry-
gian melody " clearly me ant "melod ies in the Phrygian mode," but no t neces -
sarily "melo dies orig inat ing in Phrygia ." Evidentl y this is what was meant by
"melodies of Ma rsyas" or "of Olymp us." As we have seen , the Phrygian mode
was probably very rec ognizab le as music, but it was less by virtue of being in
the Ph rygia n mode th an by virtue of originating in Phrygi a that these melo-
dies we re endowed with the power of "induc ing pos session ." In ocher wo rd s,
their effect was due less to their musical characteristics than to the fact that
the y were signs: signs of Phrygia, the land from which Dion ysus himself had
come; in short, from the cradle of Dion ysus wo rship . T his process conforms
with the gener al logic of possession which frequentl y makes a point of reveal-
ing th e fore ign or igin of the god respo nsible for tr an ce throu gh all sort s of ex-
ternal signs . Possession is essentially identific ation with anothe r. To assign
226 CHAP T ER FIV E

th at other a precise homel and is to assert his identity and thus his real ity. As
we saw in an earlier chapter, in Benin , w hen peopole belonging to Gun tribes
are po ssessed by Yoruba divinities, the y speak Yoruba and sing Yoruba
songs. In South Africa, whe n the Thong a are posse ssed by spirits of Zulu or
Ndau origin , then they sing Zulu or Ndau runes. In Greece, in order to be-
come possessed by Dion ysus , on e needed runes originating in Phrygia .
Let me sum up. Dionysiac p ossession and the Ph ryg ian mode are so
closel y asso ciated in the liter atur e of ancient Greece th at this rel ation be -
tween music and trance would seem particularly interesting to us, since it
springs from indigenous thought and attributes the onset of trance to a
specific cha r acter istic of music. In thi s ca se we would then have found-at
last - an example in w hich the power sui generis of music is explicitly held
resp ons ible for tr ance. Things woul d seem all the more explainable be -
cause ( 1) this specific music al ch aracter istic- the Ph rygia n mode - is easily
recognizable and constitutes one term of a very clear opposition ; and (2 )
this Phr ygian mode seems to h ave, unlike the Doria n- the other term of
the opposition - much greater exp re ssive reso ur ces, thus m ak ing it more
capable of producing an emot ional arousal tha t can easily be im agined as
leading to trance. A careful exam in at ion of the texts, howeve r, demon-
strates th at none of this is actually the case. Whereas the texts confirm th at
all the melodies ind uci ng tr ance we re in the Phr ygian mode , the y by no
means say that all melodies in the Phrygian mode induced trance. The
writings of Plato and Aristotle effect ively establish the first rel ation , or im-
ply it; they do not establish the second . Indeed, had they done so, the y
would have run counter to the most element ary good sense, as a reductio
ad absurdum will easily demonstrate. To asse n that the Phrygian mode
triggered tr an ce by the ope ration of some mysterious mu sical power
would have been also to asse n th at half of Greece was permanently
thrown into this state. For , as all the facts show, aulo s pla yers of both
sexes - and consequently Phrygian music - did not lim it their acti vitie s
just to Corybantic rituals. Their services we re constantly sough t p articu -
larly fo r banque ts (o r symposia ) at which there is no record wha tsoever of
possess io n ever taking place . Bot h Arisrotle and Plato, then , w isely re-
frained from asserting any such abs ur d it y. O thers, howeve r, ha ve no1
hesita te d t0 do so for them , and in their name. By recognizing, not with ·
out good cause, that mus ic p ossesses grea t imitati ve powers, and by at ·
tribut in g ethos endowed with powerful effects to th e modes, Arisro tlc
ope ned the p ath t0 all sorts of exaggerated theories . During the Renais
sance, in their wild admiration for antiquit y, people hurled them selve :
down this path in an abe rr ant way. And it is tO these aberrations, survival
of which are still discern ible roday, t hat I shall now turn.
 6 The Renaissance and Opera

FROM MARSYAS TO CLAUDE LE JEUNE; OR, THE E FFECTS OF MUSIC

Let us now leave Plato and Aristotle in order to move on rwo thousand
years, or almost, to the Renaissance, when in Ital y, as in France, Greece
and the Greeks were being rediscovered. Ir is a Greece that had been
strangely transformed, no doubt, over the centuries, in Alexandr ia first,
then Rome, then by the early Church Fathers and their successors, but one
in which we st ill find our principal themes : first "enthusiasm" o r trance
and possession, and second the ethos of modes and the Dorian / Phr yg ian
opposition; in other words, the effects of music.
The Renaissance occupies an important place in the history of the rela-
tions between music and possession for two reasons. First, it is the
wellspring, unless I am much mistaken, of the ideas some people still hold
nowadays about these relations; chis is what we shall first try to demon -
strate. Second, it was during the Renaissance that opera had its birth, as
we all know; and in my opinion opera is nothing other, in many respects,
than one of the avatars of possession. For in opera possession reali zes one
of its essential aspects, namely the identific ation of the subject with the
hero by the combined means of music and drama. As Michel Leiris has so
clearly shown, possession is fundamentally theatrical. Reciprocally, op -
era, as theater, is a form of possession. But this is a vast subject, and one
that would lead us too far astra y. I shal l therefore limit myself to a brief
sketch of the manner in which the problem can be raised in a short section
at the end of this chapter.
Be.fore coming to the Renaissance, however, let us briefly consider some
of the Latin and Greek writings that form milestones, as it were, along the
path taking us from Marsyas to Claude Le Jeune, from Pythagoras ro
Mersenne, and from Plato to Pontus de T ya rd. Our Ariadne's clew will be
the Phrygian mode we examined at such length in the last chapter, on the
one h and, and, on the other, a certain story about music and trance that
will illustrate its effect in a very significant way .
Although he lends the greatest importance to what Aristotle later
termed the ethos of modes, Plato, as we have just seen, never posited the
slightest relation between the Phrygian mode and trance. Aristotle, on the
contrar y, did so, but in a rathe r nuanced manner. Let us say that, for him,
the relation between the Phrygian mode and "enthusiasm" clearly exists,

l
228 CHAPTER SIX

but that it nevertheless admits of a fairly large margin of indetermination.

With Plutarch, four hundred years later, the final step has been taken: this
relation is raised to the status of a rule . In his Erotikos (758 - 59 ), Plutarch
cites the theor y of the four aspects of mania, as it is set forth in Plato's
Phaedrus. He then adds an obse rva tion that is of particular interest to us
on "the furor termed martial or bellicose ," about w hich he says: "eve ryo ne
knows which god sent it and provoked its Bacchic transports." 1 Then,
having pointed out that "the soldie r, once he has laid down his arms, sheds
his warlike madness" (polemike mania ), he adds: "In the same way, for
Bacchic and Corybantic leaping dances, it suff:ces tO change the rh ythm by
abandoning the trochee, and the air by abandoning the Phr ygian mode,
for them to lose their violence and draw w an end."' If in order for trance
to cease all that is necessar y is to abandon the trochee and the Phr ygian
mode, then no doubt intoning them is all that is required for inducing it.
We see that the relation between mania on the one hand and the Phr ygian
mode and trochaic rh ythm on the other, is seen as resulting from a quasi-
mech anic al effect of music.
Fift y yea rs earlier-since it was wr itt en down by Dion Chrysosrom 3 -
there was an anecdote in circulat ion that one could see as a cusrom-built
illustr ation of Plutarch 's rule . I shall not give it in this version, to which I
did nor have access, but in anothe r one, wr itten down three centuries later, in
the fourth centur y, by a Chu rch Father, Saint Basil, archbiship of Cappa-
docia. In recountin g this anecdote, he wished tO demonstrate the necessit y
for study and exercise in order to de velop the body and mind. Having first
compared athletic training to the study of music, and having poked fun
(I'm not quite sure why ) at Marsyas and Ol ympus, he praises Tim othy,' "a
m as ter in the art of manipulating sounds. So much so that he was able at
will, simply by the power of his music, either to arouse or t0 suppress
men's ardor ." He adds: "it is sa id that one d ay when he was playing his au -
tos in the Phr ygian mode before Alex ande r, during a banquet, he caused
him to rise and rush to his weapons; then caused him to return t0 his
guests agai n, thanks to a rel axed harmom·.'·
The fact chat Alexander was well known for his irascible temperament
does not alter the facts : chis an ecdote quite pl ai nly describes a trance, a
ma rti al "furor" such as Plutarch describe s with reference to "enthusiasm.'·
le is as if one were reading Michel Leiris's L'Afrique fantome, and seeing
Malkam Ayyahu being suddenly possessed by a warlike ziir upon hearing
a gunshot. Indeed, Alexander's behavior has all the abruptness and all the
conventional aspec ts of possession. The sudden alienation chat invade !
him is de vo id of all logical reason. le is not caused by anger. It has no moti-
va ting cause other than che music. Such beha vior is evidently similar tc
that described by Plato in the Ion (536c ). when he obse rves chat th ose whc
are corybantizing "have a sha rp ear for one tune only, che one which be -
The Renaissance and Opera 229
longs to the god by whom they are possessed, and to that tune they re-
spond freely with gesture and speech , w hile they ign ore all others." The
Corybantes beh av e like autom ato ns, and it is indeed like an automaton
that Alexander beh aves. The only difference is that in the case of the Cory -
bante s there is overt possession , whe re as such is not the case with Alex-
ander. T he music acts by means of its ow n power alone, wi th out the aid of
any ritual, simp ly by th e effect of the Phryg ian mode in th e first pl ace, then
by a change of mode.
Wes-hall find that this anecdote resurfaces in the Renaissance, in almost
identical form, but in a modern context. Before coming to this, however,
let us turn to another story, different, bur derived from the same m odel,
and significant in another way. It is even more inter est ing since it appears
in writ in gs even earlier than th ose cited fo r the previous exa mple , notably
in Cicero . T here are a number of va ri ants; 5 I shall use the one by Boethius,
wh ich is the most complete. It appears in the introduction to book 5 of the
De institutione musica ( 1867, 184-85 ), wh ich deals with the moral effects
of music . Boethius asks, "Is there anyone who does not know the sto ry
abo ut the yo ung man from T ao rmina who, d run k and exci ted b y the
sound of the hypophrygian mode, was calmed by Pvthagoras w ho
brought him back to his senses by changing the mu sic ro a spondee?" And
he explains: "Whe reas a prostitute had been locked up in the house of a ri-
val and whereas he [the yo ung Taorminian], in a frenzy, wa nted to burn
the place dmvn, Pythagoras, who was wa tching the night sky as usual and
obsen-ing the course of the stars, having realized th at ex cited as he was by
the Phrygian mode,6 he [the young man ] would refu se, despite the m any
warnings of his friends, to change his mind, advised that the mode be
changed, thus temperin g the frenz y in the young man 's soul and [resto ring
him ] to a peaceful state of mind." Less glo ri ously but just as nai vely stat ed,
this is re ally the story of Tim ot hy and Alexande r all ov er again. C icero's
version tells us that this Phr ygian music was played on the tibia, in othe r
words on the aulos,; which could not be more in keeping with a context of
drunkenness and orgy. Alexander was seated at a banquet. Th e yo ung
T aorminian was drunk . Alt hough there is no question of possession, both
contexts are ne vert hel ess Bacchic, in the vulgar sense of the term, if not ac-
tually Dionysiac. But thi s time, and this is just as important, we are also in
the context of Pythagoreanism, or, more probabl y, that of Neopyrhago r-
ean ism (since nothing proves that Pythago ras was eve r really the lead in g
figure in so ludicrous an adventure ), and thi s merits a closer look .
Pythag o ras thus advises that the music 's mode be cha nged, and th at the
Phrygian be repl ace d by a spondee; in other wo rds, he o rders the autos
player to sw itc h to the Dori an mode. 8 We thus are deal ing wi th a classic
example of t he ethos of modes. But the presence of Pythagoras lends it a
quire different dimension. 9 In the domain of mu sic, Pythago ras repre-
 230 CHAPTER SIX

sents, as we know, both the theory of consonance and that of the "har -
mon y of the hea venly bodies," 10 the two being closely allied. And the Bo-
ethius version tells us that at the time of the incident Pythagoras wa~
observing the stars; so it really is the Pythagora s-of-the -harm_ony-of -the-
he avenly -bodies who is involved. This version of the anecdote therefore
bears upon Pythagorean theor y as well as that of the ethos of modes.
The harmony of the hea venly bodies (though we really ought to say the
"scale" of the heavenl y bodies, since this is what the words really meant )''
is originally, let us recall, the idea that the law of numbers, which governs
the consonance of the principal intervals (octave, fifth, fo urth ) of the
scale, also governs the relations between the distances of the planets. It is
the mystery of the tetraktys, that is to say of the four first numbers which,
through the series of ratios ' 2 1 /2, 2 / 3, and 3 I 4, engenders both the con-
sonances of octave, fifth, and fourth, and the harmony of the universe. It
is also the idea that the seven strings of the lyre correspond to the seven
planets; a number obtained by adding to the five planets proper the two lu-
minous bodies, moon and sun. This theor y, however, was merel y germi -
nating in Pythagoras' discover y of the relations of consonance, he himself
never formulated it. It was Plato who did so, without referring to Pyrhagorai
moreover, and who in the Tima eus gave it its final, grandiose, and esoteric
expression (the relation to music is not at all obvious ), and thus ensured iti
fame. The theory was then taken over and modified by many writers, Cic-
ero and Boethius ' 3 among others. So behind the na'ive and ludicrous stor)
of the yo ung Ta orminian, drunk, excited, and then calmed down by th,
aulos - a degraded image of the relations between music and trance-w,
find not only Pythagoras but also Plato, and with them what was later t<
be called "the music of the spheres." Of all the Neopythagoreans wh,
wrote on music, Boethius is certainl y one of those who most influenced th
ideas held during the Renaissance about its effects. It is now time to turn t•
them.

''i There has never been a period when musicians, poets, and "humar
;i .,·_,·,.:
I.
ists"-all united at this time by one and the same desire, that of revivin
;,! '
f
Greek art-ha ve been so deepl y concerned about the power of music ;;
the Renaissance . In Italy, as in France , the main concern was to redisco v{
the secret power that music was believed to ha ve had in antiquity. Peop :
referred , with utmost seriousness, to the story of Orpheus taming the w il
beasts with the spell of his songs and to that of Amphion displacing srorn
I
I: with the sound of his lyre to build the walls of Thebes. Even the kings p1
their word in. Like the emperors of ancient China, they held music respo1
sible for public morality. In his Lettr es pa tentes, '" which created the Ac
demie de Poesie et de Musique in 1570 in Paris, Charles IX declared: "It
of great importance for the morals of the citizens of a town that the mus
The Renaissance and Opera 231
current in the countr y should be kept under certain laws, all the more so
because men conform themsel ves to music and regulate their behavior ac-
cordingl y; so that whenever music is disordered, morals are also de-
pr aved, and whenever it is well ordered, men are well tutored." Created to
"restore the usage of music in its perfection," that is to say, in the sryle of
antiquity, this Academie was intended to make its "Listeners ... capable
of higher knowledge, after the y had been purged of w hate ver ma y remain
in them of barbarity." One might think one is reading the Li Ki (Memorial
of the Rites ) of the ancient Chinese. In fact, via Florence and the work of
Ficino, ad apted to French tas te by Pon tu s de T ya rd , wha t is being re vived
here is Plato 's Laws and T he Republic. And in the "purged" we just read,
we ma y also recogni ze Aristotle's catharsis fro m the Politics. With these
Lettres patentes, which merely repe at the terms of the letter in which Baif
requested the creation of the Academie by the king , the whole program of
the Plei ade poets was given official recognition. Throu gh them, Ronsard's
very theor y of the union of poetr y and music was made state policy; hence
their importance.
Intim ately assoc iated with poetr y, mus ic was seen in the Renaissance as
a decisi ve factor of civilization. It raises the sou l and refines manners. By
wha t means? E ssentially by its capacity to arouse emotions. Among the
ancient s , Pontus de T ya rd w rite s in his Solit aire second ou discours de la
musique, "music served as an exercise to temper the soul into a perfect
condition of goodness and virtue, arousing and cal ming , by its native
power an d its sec ret ener gy, the pass ions and affecti ons, while its sounds
were borne from the ears to the spiritual parts ." Such is the model to be
followed . T hanks to the intim ate union of wo rd s and melod y, of "me a-
sured ve rses" and mu sic , also "me asu red ," it is possible to obta in the three
desired "effec t s," which are , Ba'if writes in h is letter to Charles IX, to
"tighten, unt ighten, and calm men's minds." Th e sup reme aim of music is
thus clearly defined: it should move us . And the re sult of these "eff ects" is
described fo r us by Ro nsa rd ' i in his hy mn to th P ca rdina l of Lorraine, a
poem in w hich , in prai sing the cardinal's mu sic ian, ca lled Fer abos co, the
poet also praises what he sees as the ideal in the field of mus ic:

Oh hea\·ens, what sweetness, what ease and what pleasure

The soul receives when it feels itself seized
By the movement, the sound, and the voice combined
Th at your Ferabosco on three lyres conjoins
When the three Apollos , singing divinely
And gently wedding the lyre to the voice
Suddenly, with agile throat and hand
Make Dido die again through Virgil's verse
Almost dying themselves; or \\·ith loude r trill
Rethunder the sieges of Calais or Guienne
232 CHAPT ER SIX

Your brother's victories. T hus the re is no soul

That does not leave its body and swoon away
At their sweet song, just as, up there in the sk ies
At the god Apollo's song, the gods all swoon
As he plays his lyre ....

T he ideal , as we see, is that mu sic should make one sw oon. And w hat is
swooning if not falling int o ecstasy or trance, in short to be beside oneself?
"There is no so ul that does not leave its body." Again we are reminded of
that great coming and go ing of sou ls we encountered earl ier ' 6 in H aiti and
Africa.
Exaggeration , one ma y say. A mere figure of speech th at should not be
taken seriously. Nor at a!JI The lite r atu re of the time is replete w ith refer -
ences to/ureur and furore, words that tr anslate the Greek mania, and that
altern ate w ith "enth usiasm ." In his H istoire et chronique de Provence
(1614, 583E ), Nostradamus (not the as trologer but his son ) w rites that
the troubadours der ived their poetic invention from "a certain insp ir ation
and d ivine frenzy called enthusiasm by the Greeks ." A centur y later , in his
Le Parna sse ou l'Apotheose de Corelli, Fr an<;:ois Coupe rin was to express
in music "the enthusiasm of Corelli caused bv the wa ter s of Hipp ocrene ."
H ow are we to explain this r athe r unexpected resurgence of "e nthu si-
as m ?" Largel y by the influence of two "humanists" somew hat forgotten to -
d av, but who in the ir time pla yed a dec isi\e role in this return to antiquit y
that so profoundl y m arked artistic life in the Ren aissance: Ficin o, the
great tran slator of Plato , and then Pontu s de T ya rd, 11 the masterthinker ,
so to speak, of the Plei ade . T ya rd 's book Solitaire premier, ou Discours de
Mus es et de la f ureur poetique was published in 1552 and reprinted in
157 5, a nd its title evidentlv echoes that of Ficino 's co mmentaries on th e
Ion " published earlier unde r the subt itle "ve l furore poetico."
T his "poetic furor" is nothing other than Plato's poe t ic mania, wh ich
was attributable, as we sa\,· , to the Muses. Both Fic ino and T ya rd resusci -
tated the Pl atonic th eory of the four aspects of mania, w hich , via Latin ,
became f urore or fu reu r. In resuscitating it, howeve r , the y also trans -
formed it somewha t. ' 9 T imes had changed since Plato . Of the four aspects
of mania, three had become r at her difficult to celeb rate : Fir st , ero tic ma -
nia, totally identified w ith pederast y b y Socrates; seco nd, prophet ic ma -
nia, whic h had acquired a wh iff of fire and brimstone; last, te lestic mania,
that of possession, which of course was no longer ove rtl y practiced, under
the penalty of being brought before th e In quisition . Of the se three, how -
ever, t he las t was to su rvive the best, ow ing to a cert ain amount of clever
juggling that enabled Bacchus , w ith th e help of communion wi ne, to be -
come identified wi th J esus, and Di onysianism w ith religio u s fer vo r. ' 0 But
the onl y one that had remained at all eas ilv defe nsible was poet ic mania,
w ith the result that it eclipse d the other three, and "poetic furor " came to
The Renaissance and Opera 233
be t he only st ate re pr esentative of "e nt husias m ." Let the re be no mi st a ke
howeve r : th is poe ti c "furo r " was srill very close to religious frenzy, and
was regarded as an inspiration, in rhe religious sense of rhe rerm, as a visit
from rhe spi r it; in other words, as a trance sta re, if not as possessio n .
Let us now return to the "effects" of music. If in the Dorian mode, music
will incite men to mode r ation and virtue, thus exerting rhe moral influ -
ence referred to in rhe L ettres patentes. If Ph ryg ian, ir will unleash the pas -
sions, enthusiasm, and ultimately violence . Here, in a much simplified
form, we recognize the theor y of the ethos of modes as set forth by Ar is-
totle. The poets and musici ans of rhe Pleiade took it very seriously indeed.
" If it please God to be a ble by rhe Dorian mode to extinguish rhe furor that
the Phr ygian ma v ha ve aroused," Claude Le Jeune writes in the dedication
of his D odecacorde to T urenne. ' 1 Fort y years later, Marin Mersenne h im -
self, having first expressed the opinion, in his Harm onie universe/le," that
"bad music [mea n ing Phr ygian ] should be banned from society," w ishes
rhat "magistrates would institute pr izes and rewards for those who prac-
tice no ne but Dorian music. " In that time of religious wars it was appar-
ently the secret hope of Ba'if that the psalms in "measured verses " would
disarm the rebellious Huguenots a nd that the pacif ying effect of the Dori -
a n mode would restore harmony between Catholics a nd Protestants. 23
The least we can say is that Renaissance man took music seriously'
othing better illustrates how these "effects" of music were regarded in
their relation to the ethos of mode s, on the one hand, and to ·'furor" on the
other, than two anecdotes that relate events of the time but are based on
the model that had pre viously inspired rhe stories about Timothy and
Alexander and a bout Pythagoras and the young Taorminian. Both these
Renaissance anecdotes follow the same basic scenario: first, music that
renders yo u beside yourself; second, music that restores yo u to yourse lf. It
is quite eYident that rhe y are both , in fact, simply ne\,. ava tars of the sto -
rie s related by Saint Basil and by Boethius some thousand yea rs ea r lier.
The y do differ in one respect, nevertheless, in that the first falls into the ec-
st a tic tr adit ion, the second into that of trance.
The first \,·e O\\·e to Pontus de T ya rd (1555, 113 - 15), \\'ho a fter quot -
ing the ve ry tw o anecdotes already mentioned, that of Pythagoras and the
voung T aor mini an, then th at of Tim othy and Alex ande r , passes on to a
story told to him at first h and by a contemporary, rhe count of Vintimiglia.
The count rells how, in Milan, \,·here he had been im·ited to "a sumptuous
and magnificent feast, held in honor of rh ar city's most illustrious com -
pany," he heard a lure pl aye r \, ·ho had ·'so divine a fashion of touching"
and making "the strings die beneath his fingers," rh a t all those who hea rd
him "rem ained deprived of all feeling, apart from th at of hearing, as if rhe
soul, having deserted a ll rhe sears of sensibility, had ,virhdra\,·n to the edge
of rhe ears, so as to take its pleasure more easily in so ravishing a svmphony ."
234 CHAPTER SIX

And the count of Vintimiglia adds: "st ill had we been the re, if this sam e
man [the musici an], changing his mind I know not ho w, had not revived .
hi s stri ngs, and little by little invigo r atin g his play with gentle force, sent
back ou r soul and feelings to the pl ace whence he had drawn them: not
without leaving great astonish ment in us all, that we had reco ver ed from
an ecstatic transport of some divine furor. "
Doubtless one must allow for a cert ain amo unt of liter ary license in this
sto ry; but it would be an erro r to see it as no more than hyperbole and mis -
use of la nguage. The em ot ion experienced was so great, we are told, that
it was of the order of an "ecst atic transport." "Such power is very cert ain,"
Pontus de Tya rd adds with respect to this mus ic, "and I can m yself bear
witness to such things." If he insists so much , it must be because the facts
are seen as hardl y belie va ble. So extraordinar y, in fact, that Mersenne 24 in
turn was to use them as pro of of the power that "meas ured music" in the
ant ique style could exert. T yard also refers to "divine furor," or in other
words enthusiasm. Here, certainly, we are in the realm of tr ance.
The second story is more significa nt still. We owe it to Art us Thomas, "
Sieur d'Embry, author of a volu me of Commentaires on the life of Apolloni us
of T ya nus, the famous first -century eop yt hago rean . It takes place in
1581. Sumptuous festivities had been organ ized for the m arri age of the
due de Joyeuse, favorite of Henri III. Th e Ballet comique de la R eine by
Beaujo yeul x- a pr efiguration of op er a- was presented at the royal court
fo r the first time. There was also music contributed by Claude Le Jeune,
the musician most closely associated with the Pleiade, along with J acques
Mauduit . As though to authent icate the scory he is about co relate, Th omas,
like T ya rd, begins by referring to t ha t of Tim oth y and Alexa nder , takin g
care to mention th at it was through the Phr ygian mode, then throu gh the
"sub-Phr yg ian," tha t Alexander was first aroused co furor and then re-
sto red to him self. He then cont inu es:
I have seve ral rime s heard the Sieur Claud in Le Jeune tell . . that there
was once sung an air (that he had composed along wirh the parts ) at the
magnificences provided for the wedding feast of rhe late due de
Joyeuse ... which air, being tried in a concert played fo r rhar pu rpose,
caused a gentleman present ro grab his weapons and swear aloud that it
would be impossible ro pre vent him from going out ro do battle against
someone; and rhar then they began ro sing another air'• in rhe sub-
Phrygian mode that m ade him as tranquil as before : an event rh ar has
been further confirmed ro me since by several who \,·ere present , so
much do rhe modulation, the mo vement , and rhe conduct of the voice,
conjoined rogerher, hav e force and power over mens ' minds. [Thomas
1611, 281 )

This same story was repe ated a few yea rs later by Titelouze, who wrote
an account of it to Merse n ne (who was very interested in the "effects " of
The Renaissance and Opera 235
music ) in these terms: "I remember having heard the late Claudi n le Je un e,
an excellent musician, when talking of the effects of ancient music, say
that ne believed it was w ith meas ured verses, and that he himse lf, wi th
measured French verses such as Ba-if and others composed, once sent a
captain into a furor by musical movements that he had joined to app ropr i-
ately fashioned words.»,;
With this anecdo te we are once again, as we can see, in the direct line of
Plut arch's rule: changing the musical mode is enough to trigger trance or
cause it to cease. This is ho w a continually strengthening tradition arose;
accord ing to it music could induce tr ance as it does in possession rituals,
but outside the framework of any ritual and without reference to posses -
sion - in other wo rd s, w ithout context and by its own po we r alone; or
aga in , as Pomus de T yard himself wro te, by the simple play of its "native
power" an d "secret energ y."
Moreover, based as it is on the Timoth y and Alexander scena rio, which
derives from the same model as that of Pyt hagor as and the Taorminian,
the story of the due de Joyeuse's gentleman should also be placed in the
same eopythagorea n perspective as the latter. Ficin o (who naturally also
cites the Alexander story ), Pomus de T yard, and Mersenne had all read
Boethius, 2 8 and the y we re all deepl y impregnated with Pythagoreanism
and eoplatonism (a eoplatonism that so ught , moreover, ro reconcile
Plato and Aristotle ). It is this context that explains the success of o ur story,
and it is in such a context that it should be re ad: the ethos of modes, of
course, but also " furor," or in other words trance, and "music of the
spheres," which is to say the mystique of numerol ogy and as trolog y-
these are the messages concealed with in it, with that taste for secrecy so
typical of the Ren aissance, a time when knowledge and esoterism naturall y
went hand in hand. Rons ar d's Preface sur la musique, 29 which he dedi -
cated to Charles IX, provides a very significant example of this. Writing
abou t the Phrygian and D orian modes, the poet naturally cites, among
other stories, the one abo ut Tim o th y and Alexander. In the space of four
pages there are four mentions of "the harmon y of the universe" or of the
heavens, and two of "furo r"; he who is not "stolen out of himself" by "a
sweet -sounding instrument o r the sweetness of the natural voice, ... has
a twisted, vicious , and depraved soul," and he who "does not honor Music
as a small pan of that which so harmoniously (as Plaro says ) moves all this
great universe" is not "wort hy of seeing the light of the sun." Ronsard's im -
portance for the French Ren aissance lends great interest to this musical
declaration of faith. As we see, it closely associates trance or ecstasy with
the harmon y of spheres.
This said, the principal role in producing the "effects" of music, as the y
were conceived in the Renaissance, was ne vertheless assigned to the ethos
of the modes. Whether in the case of Alexander or that of the due de
 236 CHAPTER SIX

Joyeuse's gentleman, furor and the return to normalcy are attributed to

the successive effects of the Phrygian, then the H ypo phr ygian modes. In
the story of Pythagoras and the Taorminian it is the Dorian , a more classi- ·
cal view of things, that ensures the "fu riou s" you ng man's return to his
senses. All our stories are thus centered around an antithesis between two
modes that have opposite effects; the first, in all cases, is the Phr ygia n,
which remains, whatever centur y we are in, the mode of "enthusias m. " I
shall now examine the problem of the opposite effects produced by these
two modes.
To begin with, the problem was by no means the same in the Re nais-
sance as in the time of Aristotle and Plato. No fifth -century Greek could
ever confuse the Dori an and Phr ygian. The two terms designated two
quite distinct modes 30 and w hate ver each one m ay actuall y have been (a
matter to which we shall not return ), there was no ambig uit y at the wo rd
level. By the time of the Ren a issance, howe ver, this was no longer true .
The general opinion of the age was, it is true, th at the ecclesiastical modes
in use were nothing other than the ancient Greek modes, but unanimity
ended there. The greatest confusion reigned as to which was which. Ac-
cording to Glarean, the D or ian was the mode of re, and the Phrygian that
of mi; acco rding to Galileo and Doni it was the reverse, and according to
Zarlino something different again : the first was the mode of do and the
second that of re. 3 ' All were agreed, however, in mak ing the Phrygian the
mode of unreason and the Dori an that of moderation and w isdom. Pontu~
de Tyard supported Glarean's system, Mersenne that of Zarlino. As a re-
sult, Pontus de T yard was convinced chat enthusiasm was triggered by tht
mi mode an d Mersenne, sixty years lacer, chat it was triggered by the rt
mode , the ver y mode which for T ya rd preci sely had the power of calminf
it . In other words, each man attributed diametrically opposite effects cc
the same mode! The absurdity of chis situation seems to have escapee
Mersenne, an d his faith in the ethos of modes does not seem to have beer
affected in the lease.
Parado x ically, Claude Le Jeune, to whose music his contemporaries at
:!•;:\I. tribuced such powerful effects, was much more cautious in this respect
His sweeping st atements, quoted earlier, abou t the power of the Doria 1
mode to "extinguish the furors that the Phr yg ian may have aroused," wer
much more an expression of contemporary opinion in general than of hi
own. In the rest of his D ed icace, he remains very prudent indeed, and eve
writes that the "d iversity of opinion" concerning the names of the modt
,;
I.
had pre vented him from using chem in the nomencla ture of his D odecacord1
! This alone implies that he was skeptical abo ut the Dorian versus Phr ygia
opposition. And indeed , he believed much more in the ethos of rh yth1
than in that of mode . In his Prefa ce sur la musique mesuree, which intrc
duces Le Print emp s (160 3 ) ,,ve re ad the following:

l. ~L.
 The Renaissance and Opera 237
Rhythmics, on the contrary, was brought to such perfection by them (the
ancients] that they drew miraculous effects from it: moving by this
means the souls of men to such passions as they wished; which effects
they wished to represent to us in the fables of Orpheus, and of
Amphion, who softened the cruel hearts of the most savage beasts, and
gave life to the woods and stones, so as to make them move and
place them where they chose. Since then, this Rhythmics has been
so neglected that it has been completel y lost ... until Claudin Lejeune,
who was the first to embolden himself to draw this poor Rhythmics
back our of the grave within which it has so long lain , in order to make it
equal to Harmonics."

But by "Rhythmics" we must understand the rhythm of words as much as

that of music, or, in other words, the "union of music and poetry ." As we
saw, this is also what Titelouze expresses when he reports the srory about
the due de Jo yeuse's gentleman co Mersenne: " .. . and that he himself
[Claude Le Jeune ] with measured French verses ... once sent a captain
into a furor by musical movements that he had joined to appropriately fa -
shioned words."
We are once again back to our story. But this one (that of the gentleman
suddenly seized by furor upon hearing a tune by Claude Le Jeune ) d iffers
in one respect from that of Alexander, who also becomes "furious" upon
hearing Timoth y pla y : the music played by Timoth y \\"aS instrumental (he
pla yed the aulos ) whereas char of Claude Le Jeune was vopl. As we dis -
covered in the pre vious chapter, among the Greek s the music of mania
was primarily instrumental and the musical insrrunent was almost always
the aulos . In chis respect the Timoth y / Alexander stor ,· fits perfectl y inro
the tr ad iti o n. The srory o f Claude Le Jeune and the gentleman, on the con-
trary, does nor. Thi s is a significant fact, and \Ve \\·ill examine it now.
The great idea of the poets and mu sicians of the Pleiade was that in an -
tiquity music deri ved its expressive power from its intimate union with
poetr y. In other words, for them the only true music was sung mus ic, mu-
sic in which words pla yed a principal role. It is therefore only natural that
they should ha ve shown such interest in Orpheus . Ficino, Pontus de
T yard, Dorat, Ronsard, Claude Le Jeune all refer to him as the ve ry sym -
bol of the power of sung poetry. 33 Mersenne also gives him the place of
honor in his Harm onie universe/le by making him the subject of its frontis -
piece, on which he is depicted surrounded by savage beasts held by the
spell of his song.
From Orpheus - going back for a moment co anc ient Greece - sprang
what was later called Orphi sm . Without going into the very controversial
question of what exactly it was,3' I shall simply record what Boyance
(1936 , 39 ) has said on the subject, namel y that all "the activities of the
Orphics ma y be subsumed under the idea of incantation." Elsewhere

I.
238 C HAPTER SIX

(ib id., 33 ), he a lso writes tha t "th e idea of incanta t ion is tha t whi ch estab -
lishes the deep est bond be tween t he lege nd of O rpheus and th e r elig ious
practices of the Orphics ." For them, Orpheus represe nt ed above all else
the power of incantation, which is to say rhe magical efficacy of the
chanted word. We learn from Plato (Republi c 2. 364b-c ) that the various
charlatans and soothsayers who claimed Orpheus as the source of thei r
powers were nothing other than magicians, and that it was in their incan -
tations, their "epodes," that thei r p rincipal power resided. Th ese epodes,
which they sang while performing the appropriate sacrifices at the reques t
of a client, were what we now call Orphic hymns. These hy m ns in verse
form were , to a large extent at least, what the Renaissance humanists we re
thinking about when, within the framework of a general ret u rn to Greek
music, they advocated, in the name of musical efficacy, a retu rn to t he
union of poetry and music . A poem pu blished in 1609 by an It al ian friend
of Ronsard, Banolommeo Delbene, who helped organ ize Henri Ill's
Academie du Palais (an extension of the Academie de Poesie et de M usi-
que )," is very revealing in this respec t. In it he describes the "City of
Truth," an allegory placed unde r the sign of Arisrotle, the ideal m odel of
the Academie. T he fou rt h and las t temple of this city is tha t of intelligence.
Upon enter ing it, one p ene trates into the realm of "enthus ias m" and of
"poetic furor," symbol ized by the "Union of Poetry and Music." Standing
in this te m ple is a statue of Orpheus casting a spell over savage beas ts w it h
the charm of his music, and there is a space on che pedes t al for the inscrip -
t ion of an Orphic hymn.
For the poets and musicians of t he Pleiade, then, the effects of music
were governed by th ree facrors: the ethos of modes, the harmony of rhe
spheres, and the power of incantation. T he power of music is thus seen as
being at once moral, religious, and magic. It is within this general concep -
tion, therefore, that we must situ ate the effect that Claude Le Jeune's mu -
sic had upon the "furor" of our gentleman. T hose who use this anecdote as
the perfect illustration of the power of '·measured" music in the anc ient
st vle are consequently making simultaneous reference co Plato (the theory
of enthusiasm ), Aristotle (the theory of the ethos of modes ), Pythagoras
(the cheor v of che harmony of the spheres ), and co Orpheus (the pract ice of
incantation ). Whether explicit or nor, these references explain why they
rook this anecdote so seriously. For chem, it illuscraces their gene ral repre -
sentation of the relations between music and human behavior.
For us, the sror y of our gentleman has a quite different meaning . We
have said that among che Greeks possession music was principally played
on the aulos, wordless most of the time, and intended co make people
dance. Plato, when referring co such music, uses che words melody or
rhy thm, never the word ep ode , which he resen·es for completely different
contexts. For him, incantation and possession are rwo rocally dist inct do -
The Renaissance and Opera 239
mains. Incantation is used in the serv ice of magic - or shama nism, if, like
Boyan ce (19 3 6, 59 ) and Dodds (19 51, 14 7 ), one w ishes to see the Orphi c
priests, and even Orpheus himself, as shamans - but never in that of po s-
session. In this respect , the situation in ancient Greece perfectly illustrates
the rule formulated in an earlier chapter. 36 So by attributi ng our gentle -
man's trance to a kind of music that is not on ly vocal but reputedly Orphic,
and therefore in the incantatory tradition, our stor y flatl y contradicts pre -
cisely what it claims to prove, namely its conformity w ith trance music as
it ex isted in antiquity. It runs cont rary to the manner in which the relat ions
between music and possession are practiced, whether amo ng the Greeks
or elsewhere. In this respect, then, it could be sa id to be a "bad story." But
this is not important. What is important is that, by claiming to be in con -
formit y with the Greek model, and at the same time by attr ibuting the
trance to the incantatory nature of the music, it leads one to believe that
trance music among the Greeks was incantation. Thi s is where the confu -
sion lies, a confusion entirely symptomatic of the general confusion that
the furious ferment of ideas in the Renaissance created in so many areas.
A music capable of triggering trance by th e power of its "secret energy,"
as Pontus de T yard wou ld ha ve it , or by the effect of the "hidden relations"
of the modes with the soul's "affections," as Saint Augustine wrote, 3 7 or by
the pla y of the law of numbers, or by its invisible links with the harmony
of the universe, or by its incantator y powers; in short, a music that acts
mysr~riously - such, ultimately, is the theor y that underlies our anec dote.
Ir is for this reason that it is so significant, and th at I ha ve spent so much
rime examining it. The ideas it reflects are far from ha ving lost all credit in
our own da y, and this is the point to which I will nO\\' turn.
"Incantato ry airs whose effect was irresistible on certain subjects, wha t-
ever the sk ill or the mediocr it y of the musician ... "; ;s "possessed thrown
into trance by the demoniacal call of the incantator y melody"; a possessed
person who "obeys the incantat ion of this music "- all these phrases dem -
onstrate that for Jeanmaire it is the theor y of the inc antatory power of mu-
sic that explains the fundamental role it pla ys in the possession process.
Elsewhere, speaking of rhe musical theme specific to the possessing sp irit ,
Jeanmaire (1949, 470 - 71 ) tells us that "this theme particular to each of
the various demons by which each dancer believes himself haunted thus
automaticall y triggers the onset of trance in the hitherto apathetic pos -
sessed person. " Automatic behavior al re sponse, haunting , incantation -
all this adds up to a whole that constitutes, for Jeanmaire, the climate, so
to sa y, of possession among the Greeks. Yet, as we ha ve seen , there is no
mention whatsoe ver of incantation in either Plato or Aristotle! T o attr i-
bute the effect of music on trance to its incantator y pO\ver is to interpret
the evidence in a totall y arbitrar y wa y. It means introducing magic where
it does no t belong. Abo ve all, it means being quire mistaken about the for -
240 CHAPTER SIX

mal characteristics of the music concerned. The music that triggered "the
bounding dances of the Bacchantes and Cor ybantes," as Plutarch wrote, 39
was evidently stimulating, extrovert, dynamogenic . Incantatory mus ic is
by definition introverted, slowly • turning aro und upon itself, proceed ing
by fascination, and developing itself in immobilit y. To see the music that
induces trance as incantatory music is therefore a major musical miscon -
ception. Ir is amazing that Jeanmaire, despite his perspicacit y and his ad -
mirable scholarship in the field, should have made this error . T he reasons
for it are cle arly the power of the Orpheus myth on the one hand, bur also,
and above all, that of the ready -m ad e ideas generally held about music,
and ro wh ich even the best minds are not alwa ys immune. Man y others
have made and are stil l making the same error of interpretation. T he ir
excuse is that they are following a tr ad ition d aring as far back as the Re -
naissance, and that has been maintained by writers as eminent even as
Combarieu, who in his great wo rk La musiqu e et la magie de vo tes several
pages to incantation in his chapter on the ethos of modes ( 1909 , 228,233 )
and concludes that among the Greeks the latter was simpl y "a legacy from
ritu al magic."
Linforth, to cite only one more example, makes an error of the same
order. In his ver y important article on Corybantic rites in Plaro, he refers
several rimes to rhe "intoxic ating " nature of trance music," 0 and to its "in-
roxicaring spell." Spells, inc anta tions - we are still faced with the same
general conception of things, that which consists in attributing trance to a
more or less inexplicable pO\\·er sui generis in music, which acts in some
way physiologically, like a drug, like an intoxication, like a m ag ic potion.
or once does Plato mention incantation (epodai ) or inroxication in his
texts on Corybantism . Linforrh is yielding here to the same temp tation as
Jeanmaire, w hich attributes ro music the magic power of triggering pos -
session, something w hich neither the writings the Greeks nor ethnog r a-
phic data justify.
Ir is this same Renaissance eoplatonic tradit ion that has influenced -
or perhaps given direct rise ro - anorher interpret ation of ho w music is re -
l a red to trance . Bur here it is the mystique of numbers that is invo lved.
The theory of the harmony of heavenly bodies re sts, as we saw, upon
the belief, a purely intuiti ve one, th at the movement of the planers and the
relations of consonance are gove rned by the sa me law of numbers and by
the same mystery of the tetraktys (w hich one might translate as "fourth -
ness " ) that designate the orde ring of the first four numbers in accordance
with the ratios 1 / 2 , 2 / 3, and 3 / 4 (w hich would now be called superpar -
ti als ), wh ich are those of consonance. In a wo rd, it is m ystical numero-
logy. When Alain Danielou ( 197 5, 14 ) writes that all music used to induce
tr ance , always and everywhere, utilizes rh ythm based on the numbers 5,
7, or 11, he is making use , whet her he says so or not, of a theory of num -
The Renaissance and Opera 241
bers that is admittedly different from that of Pythagorism, but that is es-
sentially the same in spirit. In disco vering the laws of consonance
Pythagoras gave proof of a gre at scientific mind . In developing the theor;
of the "ha rmon y of the spheres," w hich is pure bunkum, the Pythagoreans
and Neopythagoreans were merel y dabbling in bogus science . The y,
nonetheless, were taken seriously for a long time. It seems that even
Kepler believed in the music of the spheres." But then , was he not as much
an astrologer as an astronomer?' 2

In appealing co Plaro, and in invok ing mania under the name of /ureur
poetique , the poets and musicians of rhe Pleiade renewed the ties with the
Greek trance tradition, bur their version of it- need I say?-was literary
and wate re d-down. Parallel to this arisrocratic current at the level of the
cou rt and the salon, however, wa ano ther one, popular this time , that
cominued to keep the practice of trance alive in Western Europe in much
less civilized forms . Taramism in Ital y and Spain, epidemics of Saint
Virus's dance elsewhere , demoniac possessions occurring sporad ically in
many places, and culminating later in those at Loudun - all these were
very virulent m a nifestations of trance that ha ve a greater connection wi th
relestic traditions-those of Cor yba ntes and Bacchantes -t han wit h the
poetic tradition of mania.
After a long eclipse, the latter made a new appearance, nor so man y
yea rs ago, again accompanied by music - of course - bur this time by an
international publicit y campaign as well: Beatlemania. The recipe is ap -
parently a good one, since it has been revived even more recently by yet
ano ther rock group th at labels its show "Starman ia." Of course, the wo rd
in this comexr means nothing more than "madness." Yet its use is never-
theless significant, and revea ling of a certain desire - or a certain nostal -
gia- for tr an ce. Learned tradition? Popular tr ad ition? Juncture of the
two' Thar is vvhar remains to be seen.

LETTER 0 1 A OPERA

Opera , as we saw earlier, can be regarded in certain re spe cts as the last
avatar of possession ceremonies; or the opera singer (and this comes to the
same thing ) can be reg arded as the most recem of the roles so far played by
the possessee. This would mean, all things being equal, that opera is for
the German, the Frenchman , or the Itali an what the I.Jori is for a Hausa
and hau bong is fo r a Vietnamese. If the reader doubts th at this is so, noth -
ing could be more likely co convince him, I feel, than the letter about co be
put before him. Ir was written' 3 by a yo ung erhnomusicologist from
Benin, at presem living in Paris, to a friend and colleague back in Africa. It
fell providentially inro my hands at the very moment when I was prepar - ·
 242 CHAPTER SIX

ing ro deal with this very aspect of the matter. I am reproducing it here just
as it was written, with no additions other than a few notes to clear up cer -
tain allusions in the text that would otherwise remain obscure .

M y dear Asogba,
What an adventure! I went to the Opera yesterday. I thought I'd gone
raving mad! o one had warned me, so I had no idea what I was in for:
imagine my surprise when I found myself bang in the middle of a posses-
sion ceremony! You would have thought you were in Porto -Novo, in the
Place Degue, attending the annual feast for Sakpata, or at Alada attending
the ceremonies of Ajahuto, or at Abomey for "The Grand Customs." Of
course it's not the same thing at all, that's obvious. Of course the differ-
ences are immense. Never mind! I still think that a performance at the Op-
era and a vodun ceremony in Benin are in many respects fu ndamentally
quite comparable. That's what I want ro explain to yo u now, since no one
here seems even ro suspect it, and everything else can wait.
Let's take psychology first. You will agree that the essential aspect of
this , in possession , is identification with another, the invasio n of the field
of consciousness by a person other than the one one is normally, so that
one is no longer oneself but this ocher , and so one behaves in ever y way
like him. This is precisel y what h appens on the scage during an opera. Lase
night, in Elektra, which was the opera I attended, Birgit Nilsson was no
longer Birgit Nilsson but Electra. You will say, yes, but that doesn't hap-
pen only at the opera, it's just the same in the the ater, or also in movies.
That's true to some extent. But what makes the opera a particular case,
and links it so closely with a vodun ceremon y-I almost wrote perfor-
m ance-is its relation ro music . In both cases it is in face the music that or-
ganizes the performance, gives it the structure that governs its develop -
ment, dictates the movements, regulates the alternation between tension
and relaxation. Just as a man or woman embodying a vodun takes his or
her cue from the music co dance ouc his or her identification with the pos -
sessing divinit y, so the opera singer takes his or her cue from the orchestra
in order to express in song the character he or she is enacting . It is music
that gives life ro both of them. Neither one nor the other could incarnate
his character without being constantl y supported or even carried along by
the orchestra. Their experience of their role is essentiall y a musical one ,
which is not the case in either the theater or the mo vies. You know hov.
cruci al the function of the musicians is at home, in all possession ceremo -
nies. One has the feeling that they are the ones who lead the play, that un -
. ·I der their power the possessed are but puppets receiving their orders. Al
things being equal, I get the feeling that the conductor and orchestra hav,
an exactl y comparable function in opera .
The big difference, from the point of view of its relation to music, evi
dently, is that in a vo dun ceremony the possessed dance and don't sin§
The Renaissance and Opera 243
whereas in the opera the performers - I nearly wrote possessed again -
sing and don't dance. We could say that in vodun [or bori, or hdu bong]
possession is expressed through dancing and that in opera it is expressed
through singing . Or again, that the acto r playing the vodun embodies his
character by dancing it and th at the opera performer embodies his by sing -
ing it. For he does embody it. Open the article on opera in Fasquelle's En-
cyclopedie de la musique and you will read: "It is nor the reality of the
characters on stage that is important in an opera, but the degree to which
a singer -interpreter succeeds in evoking, by embodying [and those are the
wr iter 's italics, not mine ] the inner character in the drama." I am tempted
to think that singing a part implies a ·greate r invest ment of the self in the
dram atic ac tion than when one merely acts it. ft is nor for nothing that the
French call opera "lyric theater." To use Jakobson's terms, I would say that
lyricism means that the emphasis is placed on the "addresser" of the musi -
cal message, on the "I. " The opera house, the lyric theater, is thus the place
where the first person expresses itself. Ir is the triumph of the profound ex-
pression of the self, at its most affective, most irreducibly personal level.
To express that self it is therefore important to be totally invaded by it,
and this invasion of the self probabl y never occurs more fully than in this
lyric -dramatic situat ion, in which it is permissible to live it out totall y in
public, precisel y because one embodies a character that is nor onese lf . It is
obviously the presence of the spectators th at gives full meaning to this ex -
traordinar y adventure that the opera singer lives out when sing ing on
stage.
Does this intensit y constitute trance? T o answer this question one
would need to ask the singers themsel ves. Would they or could they, an-
swer? Once themselves again, once the curtain is down and the perfor -
mance over, do the y remember what they ha ve experienced, or are they
subject to the same amnesia that our vodun adepts experience regarding
everything that happened while the y were possessed? Ultimately, the
point is to know whether the opera performer lives his role and sing s it in a
st ate of dual personality, so that he is simultaneo usly himself and the per -
son embodied, one governing and controlling the other, or whether, on
the contrary, he is solely the other and no longer himself at all, in which
case w e are dealing nor with dual personalit y properly speaking-or, as
Freud puts it, "double consciousness"-but with a personality change due
to substitut ion. It is a question that has been hotl y debated ever since
Diderot, I belie ve . Howe ver , we mustn't lose sight of the fact that opera
performers are professionals, that ch anging personalities is part of their
trade , and that this is not the case with the vodun adept, even though he
coo h as sometimes undergone a very long training period.
Let us assume then-first hypothesis - that while on stage embodying
the hero he repr esents, the opera singer is in a trance, and let us call it , in
this context , a lyric possession trance. This tr ance is quite clearl y extremely
 244 CH APTER SIX

controlled. It is not preceded by any crisis. Could chis be the onl y type of
trance displa ying chis particularit y? And if so, wo uld it be less of a trance
for chat reason? As yo u know, the trance in vodun cults is not necessaril y
preceded by a crisis. Indeed , chis is one of th e factors chat helps in distin -
guishing between possession by a lineage vodu n and po ssession by an "ex -
ogenic" vodun, to use Luc de H eu sch's term. In Senegal, in the ndop, only
ne ophy tes ha ve crises; theore tic ally the wo men w ho officiate ne ver do. So
the fac t th at it is de vo id of cr isis wo uld not be en o ugh o f a reason for see-
in g lyric possess ion tr anc e as an isol ated case w ithin the general frame -
work of possession. (Parentheticall y, crises do in faa occur quite frequently
in che oper a house, or so I'm told. H ow ever, the y cake place in che wings;
in sho re the y are not pare of the performance. In o cher wo rds they are non-
ritualized crises, or just "t antrums ," bu t chat's another problem .)
ow lee us change our hyp othesis and say chat the opera singer is not in
a tr anc e. Lyric possession would then be a fo rm of po ssess ion without
tr ance and we woul d have a type of possession char is completel y differ-
ent, ps ychologicall y speaking, from chat at wo rk in po ssess ion by a vo -
dun . In chis case, should we say char we can no lo nger spe ak o f possession
at all ? I don 't think so. Th e crucial thin g, ulcimacel y, is th at , from the spec-
tato r 's viewpoint, the entire event happen s in such a way char the ope ra
singer is seen co be trul y embodying his cha racte r, or is, in ocher words, to ·
rally possessed by chat character. Indeed, if the spec tato rs belie ve in thii
inc arn ati on, the singer is a gre at actor. We are therefore at the front ier o:
possess ion, but still within it. Or to put it more acc ur at ely (since we mu s·
not argue as though possession constitutes a fac t in and of itself), let us Sa)
chat lyric possession has enough points in com mo n w ith reli gious pos ses
sion for both to be cons ide red as belonging co one and the same ver y gen
era] state of consciousnes s.
So, if seen in rel ation co po ssess ion as a st ate of consciousness, an oper;
per for m an ce and a vodun cerem ony are appreciabl y equivalent, ho w is i
t he n chat the y are so diffe rent? Is it, ultim atel y, because one is aimed at ch,
re aliza ti on of an aes thetic need an d the ocher at chat of a religious one? Bu
is this difference a re al or an illu sory one? W hat if the aesthetic function i1
one type of so ciet y were the same as the reli gious funct ion in another? H a
art not been, for a ver y long time , rhe m ain technique of religion? It is n,
wonder that , as the religious function atrophies, art becomes an end in it
self. We h ave heard often enough th at Greek drama origin ated in the cu :
..,
I ·
of Dion ysus, chat modern Western cheater had its origins, in part at leas ,
in che Christi an Mysteries, and that J ap ane se 6 is but a the atrili za tion c
possession . I kno w not hing abo ut the histor y of opera, bur I w ill one
aga in qu ot e from Fasquelle 's E ncyclopedie de la musique (my great sourc
of inform at ion , as you can see !), in which the article on Comic Oper
states th at its origins lie in the medie va l Fete des fous , o r Feast of the M ac
The Renaissance and Opera 245
men. Weren't these "madmen" purely and simply possessed p erso ns? Un-
less I'm much mistaken, there are a great many reasons to thi nk so . . ..
But let us leave aside these historical problems about which I know n oth -
ing, and get back to the question at hand . I tend to think that the differ -
ence between a performance at the Opera and a vodun ceremony is largely
circumstantial an d dependent on the context.
I ha ve already told you the ways in which the opera seems to be closely
akin to a possession ceremon y. I would also like ro explain the ways in
which the possession ceremonies that you and I know now seem to be akin
to opera. A moment ago I mentioned Alada" 4 and the annual ceremonies
for Ajahuto. 45 Now think back to the feast for Ajagbe that we attended to-
gether a few years ago and let me tell you how I remember it. In the late
afternoon, Akplogan, 46 sumptuously clothed in a blue and gold brocade
robe, left h is house in grand procession, preceded by spear carr iers . Sea t s
had been prep ared for him and his entourage on one side of the public
square w here the d ances were to take place, and in the center there stood
an enormous kapok tree. As the y arrived, the various families of Alada
and the surrounding district, who had also come in procession, went to
greet Akplogan before taking their places nearb y. The va rious merchants
set up their stalls slightl y behind the aud ience. Little by little the square- I
almost wrote the au ditorium-filled up . Drums, rattles, and iron bells
we re gathered at their appo inted place, but no o ne see med to pay any at -
tention tO them. When the time came, ho weve r, mu sicians appea red and
began trying them out : hammer -blows on the drums to stretch their skins,
test -drumming to judge the effects, bell-calls ske tching out a rh ythmic
phrase. A certain vibration in the air that we both love so much had taken
over. As I sat in the Opera, I experienced this same impression of mu sic
awaken ing when the musicians rook their places in the pit an d began to
prelude, as it were, each performing in his own corner, one playing a few
notes of a scale, the other a brief arpe ggio. The music slowly took posses -
sion of the place, just the wa y it does back home. Unfortunately, I have no
idea wha t exactly was going on backstage, since the audie nce isn 't allowed
there, as you can imagine. But I was suddenly reminded of that wonderful
evening in Alada, to which I shall no w return . You had gone to say hello to
yo ur kinfolk, who had just arrived and settled them selves in a spot - I al-
most wrote a box - near Akplogan. You were making your soc ial rounds,
in short, just as I saw it done last night at the Opera. Since I know ver y few
people in Alada, I went for a stroll. I was intrigued by a half -open wattle
door behind w hich there seem ed t0 be a great deal of coming and going. I
slipped through it. And I was backstage! There 's no other way to put it.
You know that the Sak patasi" ' ha ve skirts worn tightl y at the waist? Two
of them w ere adjusting their skirts, busil y knotting the little fastening
strings, then swaying their hips to see whether they hung right. Furthe r
246 CHAPTER SIX

on, a woman who was also, visibly, a vodun devotee was making herself
up, helped by a companion holding a mirror. Elsewhere, a Legba' ' was ar-
ranging his necklaces, checking the fasten ings of the vase panopl y of ob- ·
jeers they always wea r, and adjusting the tilt of his straw hat. Elsewhere
aga in , a group of women was swarmi ng around a figure whose back was
turned co me and whose loincloth was being arranged. All these people
were gaily chattering, some standing, moving from one group to another,
ochers seated on small scools. Suddenly my presence was noticed, and
with cries and exclamations from all present I was gentl y but firmly shown
the door. I'd had a terrific time. I had really seen vvhat it was like backstage
and all that went on there in preparation for a performance. I had discov -
ered chat our great ceremonies in honor of the vodun are just like theatri -
cal performances in that they require a quite a bit of preparation and in-
volve a great hustle and bustle behind the scenes, during whic h the adepts,
who prepare to go into trance and to be possessed by their gods, behave
just like acto rs backstage in oche r countries.
The drums began to beat. The feast began. The ritual actions had taken
place chat morning: invocation of the ancestors, divination, offerings, li-
bations, sacrifices, pra yers. The evening would be devoted exclusively to
dance. It became dusk. First came a very calm parade, accompanied by
drumming chat rose and fell. It was as though the entire troupe had to be
presented co che audience. If I remember correctl y, Khe vioso, 49 recogniz-
able by his recade' 0 in the form of a stylized brass ax, led the way, unless it
was Legba with his enormous phallus of polished wood around his neck.
There were about ten of chem in all, all richl y costumed, the men wea ring
shore variegated skins, rather like tutus, the women in long crossover
skirts made of multicolored cloth, arms weighed down with silver brace-
lets, neck and ears bedecked with necklaces and pendants. As they mo ved
past, the audience acclaimed them, shouting out mottoes at the top of
their lungs, their vo ices mingling with those of the singers standing near
the drums. After circling che dancing area two or three times, half walk -
ing, half dancing, the procession disappeared backstage again. There fol -
lowed a fairly long pause , filled only by drumming and rather loose sing-
ing . Tight had fallen. In the dark, one could scarcely make out people's
faces, since the only source of light in the square was the gas lamp placed
besides Akplogan. Further back, the tiny lamps of the stall keepers pierced
the darkness here and there with their small yellow flickers .
Sudd enl y, Khevioso burst into the square, eyes rolli ng, a white kerchief
knotted around his head, brandishing his brass ax in threatening gestures,
running and leaping in all directions , driving back any children who hap-
pened to cross his path. Sacred furor of the god of lightning! His brass ax,
raut curve and tongue of fire, flashes in the night. Khe vioso spins around
on himself, hurtles first thi s way, then that. His eyes roll frenziedl y in his
head, white flashes shining through the darkness with a moist gleam. Hi s
The Renaissance and Opera 247
feet m ake the dust fly. Drums, rattles, and iron bells had joined in freneti -
cally, encouraging, with the greatest agitatio n , the dancer 's w ild dashes
and turns .
Meanwh ile , wit h out anyone notic ing, another dancer - excuse me , a
female di vinit y, but wh ich ?- had also t aken po ssession of the stage, mo v-
ing forward wi th tin y steps, scarce ly dancing at all, and ver y reticentl y,
wit h little sways of the hips, arms curved out in front of her to part the
panels of her pagne. Apparition full of charm an d gentleness, making the
most touching contrast w ith Khe vioso and his violent demonstrations of
virile energy. Then it was time for Legba 's entrance, as fantastical as ever,
so that one ne ver knows wh ether he's serious or jokin g. Straw h at and tutu
both dyed purple and both of a so me wha t unlikel y shape. Legba pirou -
ettes, st ops suddenly, legs apart, holds his bizarre pose; then shoots off un -
expectedly in a new d irection . Everyt hing about him is disconcerting . H is
expressions force you to laugh, but yo u're nor qu ite sure w hy. Ocher danc -
ers - I me an other gods - come on stage one after th e other. Now there is a
pantheon of five or six divinities danc ing together, each in the style befit-
ting his or her cha racter. Legba, who had va ni shed, makes a sh owy
reentry, holding his huge wooden phallus in front of him w ith both hands,
and taking over the center of the square, he performs an obscene dance
w ith utmost vigor. General but nor excessive hilarity ensues. And so the
performance continues, with entrances, exits, crowd scene s alternating
w ith solo exhibitions. All executed w ith a very great freedom of mo ve-
ment and, certainly, wi th a great deal of improvizarion. Since I was enjoy-
ing the sho w as a dilettante, ne ver ha ving been initi ated into any cult my-
self, the esoteric aspec t of these dances naturall y escaped me. And yet ,
w ith all those gods parading abo ut , we certainly we re being treated to a
tremend o us lesson in mythology. Yer one still had to know how to read it.
I recently described that evening to so me friend s- white people, of
course - here in Paris. "But wha t abo ut tr anc e in all char?" they asked .
Well, you know, I told them, come to chink of it, there was no more - bur
no less - of it th an at the Oper a . I doubt w hether someone who was reall y
a strange r to the country, and hadn't been told beforehand, would have
even suspec ted that the people he saw dancing were in a state of posses -
sion . And yet, as yo u and I both know, they we re. Th e dancers always are
in such ceremonies. Bur entr y into tr ance rakes place backstage - whethe r
it is wi th or w ith out crisis is secondary - and from then on it is scarcely ap -
parent most of the time, except by certain small signs recogni zab le onl y to
chose experienced in such matters. Perhaps there is a slight fixity to the
gaze, perhaps it is bizz arely elsewhere. T har's all, app arentl y at least. I
couldn't say what goes on in their heads.
As I told you, so far I've seen on ly one opera. T har's not much of a basis
on which to make valid comparisons wi th our vodun feasts . The d iffer-
ences from op era to opera are very great, it seems, because the reperto ire
248 CHAPTER SIX

is vast. But that 's true of our ceremonies too! Very often, I'm told, ope ras
invol ve a certain magical element. The action can move outside everyday
rea lit y. Things happen that are scarcely believable. To limit myself to
Mozart, I'm cold this is true of two of his operas: The Magic Flute, which
as its title indicates has a fairytale atmosphere, and Don Giovanni wh ich,
toward the end, also brings into play a fabulous character, the Commen -
datore. And this supernatural, fabulous element is also at work in our
ceremonies. I would even think that it is essential. Remember that ceremo-
ny for the resurrection of a Sakpatasi, at Porto - ovo? What could be
more fabulous, more unbelievable than a resurrection? Bue what theater !
What staging! The open space at the center of the square black with peo-
ple, the macs so carefully arranged on the ground, the high priest, all ir
wh ite, sitting at one end with cwo pots full of magic water at his feet, the
corpse swathed in its shroud, making its entry on the square borne rigic
on outstretched arms, sec down on the macs before the high priest, cher
the priest's seven calls, summoning, in dramatic tones, the body to wake n .
the truly "death ly" silence of chat great crowd awaiting the miracle , the ex-
plosion of joy, the delirium chat sweeps through it at the first sign of re-
turning life, the frenzy of the drums, suddenly liberated again, the who!(
thing preceded by a slow setting of che scene, by choruses, dances, proces-
sions, and ending in a general dance accompan ying the resuscit ated Sak
pacasi 's first seeps, che songs, che joyous sound of all rhe instrument s. Anc
even though the ceremon y was very much part of a possession cult, sine(
the corpse was that of a girl killed by the vodun who wanted her for hi :
wife, there was in face no quest ion of trance. Yee possession certainly lai
at the very heart of chis drama. Dead ro her former life, the girl was bein1
reborn to ano ther existence, chat of the w ife of a god, a w ife who m chi:
god would henceforth ritually possess at each of his feasts . A change o
personalit y, through possession, was the very substance of the ceremony
and the whole thing was presented , I now realize impregnated as I an
since yesterday with opera , as a musical drama. Ir is impossible ro con
ceive of it \\·irhout song, \\·irhour dance , and without musical instruments.
Of all the vodun ceremonies I have seen back home, chis is the one tha
provides che best example of dramatization and cheatrilization . le is sig
nificant, it seems to me, chat it had as its theme chis resurrection, takini
place within the framework of a possession cult and consisting, ulti
mately, in the birth of an actres s.
Ther e is another good example of cheacrili za rion among our ceremo
nies : che "G r and Cusroms" of Abome y, which also involve a very elabo
r ace theatrical production . 5 1 Imposing processions, tremendous choruse
as they move from the squa re to the temple, and all very similar, if I ca1
believe what I'm told, ro what one sees in Boris Godunov. Long, unbroke1
succession of dances consisting of quite varied allegorical and symbolic fig
The Renaissance and Opera 249
ures whose deepest meaning is not reall y underscood except by initiates.
There are dramatic episodes recalling the grea t feats of the princes of the
blood and warriors of former kings, redoubtable fighters who decapitated
their enemies w ithout pity, an d w ho are still embodied today by entraced
wo men in the course of these festivities. At a specified moment in the cere-
mon y- w hich lasts almost an entire day and is repe ated on several succes-
sive days for weeks on end - each dan cer (although she is not the woman
one sees; but the char ac ter by whom she is possessed an d who "da nces on
her head" ) in turn bre aks away from her gro up and simu lates such a com-
bat in the cent er of the squ are. Thi s t akes the form of a saber d anc e in
which the m ovement s are precise , lightning swift , but very sty lized, pre -
ceded by a dance circling the square and followed by a return of the dancer
to her group supp orted on the shoulders of two of her companions, w ho
move forward to help our the exhausted wa rri or. All acti on is accomp a-
nied , of course, by choruses and drums. H ere, however, contrar y to wha t
we saw in Alada, there is no impro visa tion. The ent ire performance is as
strictl y regulated as an opera -ballet would h ave been, I am told, in the
days of Louis X IV.
I expec t you' re wonde rin g by now what I'm hoping to prove w ith all
these comparisons, wh ich yo u ma y think ver y supe rfici al. Don' t worry,
I'm not going to sta rt arguing that performances at the Paris Opera have
the ir origins in cerem onies for the vodun in Alada, or anywhere else .
What stri kes me is th at possession ult im atel y seem s to occupy fa irl y co m-
parable positions in two st ates of civilization as far remo ved from one an-
other as the one in Paris , on the one hand, and Porto-No vo on the othe r.
But despite appearances, or rather ready -made ideas, this is in fac t the
case . E veryo ne willingly agrees that possession occupies a very important
position in the traditional religion of our countr y ; but you will find pre -
cious few people ready co accept the idea tha t possession also occupies an
important position in the soc ial life of contemporar y France. The Opera
stands in the ve ry heart of Paris; its budget is more than a hundred million
new francs a year (a tenth of our entire n at ion al budget 1) and this is whe re
foreign heads of state are taken when the y come on official visits. If my
an alysis is correct, therefore, thi s T heatre Na ti onal de !'Opera of th eirs is
nothing othe r than the French temple of lyric possession. Nonreligious,
wholly profane possession, but possession all the same, an d comparable
to that observed among us, since once aga in opera consists in people em -
bodying, or behaYing so that othe rs believe the y are embod ving, im ag i-
na ry charac ters with whom the y identify in public by me ans of beh av ior
closely associated with music. You ma y tell me that none of this is reall y
important at the Opera, that what matters is to show off one's dres s or
one's wife, to see if the performers sing well and if the sets are beautiful.
But all things being equal, the sa me is true at home! At home, possession
250 CHAPTER SIX

takes place in the course of a performance very comparable to an opera.

An opera performed here involves behavior very comparable to posses-
sion. The fact that possession among us is totally sincere (in theory at
least ), and that it may be totally insincere here (but can it ever be? ), does
not change an ything at all.
The fascination that possession exerts on the minds of men is what in-
trigues me. I had na ·ively thought it was hardly ever found outside of, let
us say, "archaic" societies, such as our own. M y visit to the opera has
shown me I was quite wrong . How is it that this cultural trait is apparently
a uni versal in human societies, whatever the stage of their development?
To w hat deep need does it correspond? What fund amental function does
it perform? Could it be the kath a rsis Aristotle wrote about? This, I must
say, is what I was tempted to think when I heard (to my great surprise! ) the
prodigious outpouring of enthusiasm that exploded in the auditorium as
soon as the last note had been sung. One would never find anything like
that at home , I thought co myself. That unending applause, those yells of
appro val intrigued me a great deal. Were the y a manifestation of collecti ve
fren zy ?
As you ma y ha ve gathered, I still must clear up a lot of points. Ethnolo-
gists here in Europe are beginning to ask themsel ves (it is high time! ) what
the "indigenes " chink of their interpretation s. So I hope co find a few nati ve
Parisians w ho will cell me what the y chink of che ideas I've had and the
questions char came into my head as I observed chem at the opera. I'll
write again and tell you what I find out; it should be interesting. But I
shan 't necessaril y take what the y say at face value, you can count on that.
Wh y should the mere fact that they are natives place them ipso f acto in a
better position than us co understand what is going on among them?
"First came a very calm parade, acco mp anied by drumming that rose and fell" (p. 246 ).

the women in long crossover skirts made of multicolored cloth, arms weighed dow n
with silve r bracelets, neck and ears bedecked with necklaces and pendants" (p . 246 ).
"I doubt whether someo ne who was really
a strange r ro the country, and hadn' t been
raid beforehand, would have eYen suspected
that the people he saw dancing we re in a
state of possession" (p. 24 7 ).
"Legba pirouenes. . . Everything about him is dis-
conce rrins( (p . 247 ).
"Suddenly, Khevioso burst into the
square ... , a white kerchief knot -
ted around his head, brandishing
his brass ax in threatening gestures,
running and leaping in all direc -
tions. . . Sacred furor of the god
of lightning! . . . His eyes ro ll fren -
ziedly in his head, white flashes
through the darkness with a moist
gleam. His feet make the dust fly.
Drums, rattles, and iron bells had
joined in frenetically, encou raging,
wi th the greatest agitation, the
dancer 's wild dashes and turns"
(pp. 246 - 47 ).
7 Music and Trance among
the Arabs

Of all the peoples in the world, the Arabs ' are undoubtedl y those who
have associated music and trance the most closely: first, in their religious
life, w ith Sufism,' in which trance (wajd ), which for many adepts occupies
a very large place in the search for God, is achieved very often through
music; and second, in profane life, in whi ch mu sica1 emotion (tarab ) tra -
ditionall y leads, very frequently, to trance behavior. 3 These two aspects of
the relations between music and trance among the Arabs are w hat we
must now examine.

RELIGIOUS T RANCE

Kitb adab al-sama cy wa al-wajd, "Book of the Right Usages of Audition

and Trance ," 4 - such is the title given by Ghazzali (al-Ghazzali, known as
Algazel during the Middle Ages ) to the eighth section of his renowned
Ihyac ulum al-din, "Book of the Revivifying of the Sciences of the Faith ,"
which dates from the earl y yea rs of our twelfth century and constitutes one
of the most important writings on Sufism. "Audi tio n ," let it be clear from
the start, means aud iti on of what we have here decided to call "m usic." To
be sure, one could not h ave found a book more closel y connected with the
problem of the relations between music and trance .' But before coming to
it, let us first examine the wo rd s sama c and wad), which I have translated
as "audition" and "trance," whereas Duncan B. MacDonald, to whom we
owe the English translation of thi s wo rk, most often renders them by the
terms "listening" and "ecstas y."

The Word Sama c

Given the conte xt in w hich Ghazzali uses it, the word sama c is in fact un -
translarable, 6 for two reasons. The first is that it denotes a very particular
thing, peculiar to Sufism, which is the ceremon y made up of pra yer, mu -
sic, and dance that brings dervishes together for the purpose of adoring
God and pr acticing trance. This ceremon y is called the sama c and, taken in
this sense, it is similar to the wo rd is/am in th at it has no equivalent in
French o r En glish. The second reason is that Ghazzali's book is abov e all
 256 CHAPTER SEVEN

an attempt to justify sama ' and his justification rests upon the ambiguity
of the word, or rather upon the fact rh ar the word has two meanings, a
very particular one, which I have just given, and a general one, wh ich ii
"listening" or "audition." If "audit ion " (sama ') - "of poetry and music" is
understood-is lawful in the eyes of the faith because it can cite illustriou s
precedents in its support, then the sama ' (ceremony ) is also lawful.' This ,
in extremely simplified forms, is his argument. 1 ow we must look at rhe
meaning, or rather rhe meanings, of sama " even more closely. 8
The relation between sama ' as ceremony and trance (wajd ) is so close
that the word sama ' , in thi s sense, also sign ifies the trance sta re. One can
in fact say "to go into sama ' " or "to be seized by sama ' ," or again, "to be in
that state. " 9 Mole' writes ( 1963; 148 ) that the term is practically equivalent
to "ecstatic dance," At the beginning of the first chapter of his book (Mac-
Donald 1901; 200 ), Ghazzali says, "know that the sama ' . . . bears as fruit
a state in the heart that is called trance [wajd] ." Taken to its limits, sama '
can also mean "music," or some thing very like it , since one can say 10 "hea r
sama ' "- which is surprising since it runs counter to the very logic of the
word, for such a usage amounts to the same thing as saying "hearing rhe
audition." Bur, once again, sama ' , which derives from the root s.m. ' .
(hear, listen ) den ores, in its first and genera l sense, the act of hearing or lis-
tening, without reference to any particular acous tic phenomenon, or , let
us say, "audition." Granting this, in the Sufi texts that interest us here, the
verb "hear" (s .m .') in its various fo rm s always includes an implicit object,
which is either poetr y, the Koran, or music , though a particular kind of
music only. But which one? In order to answer this, I must first digress on
the wo rd "music ."
The Arabic word musiqi - which comes directl y from the Greek, need I
say? - does not appear even once in Ghazzali's book. 11 This is clearl y not
because he does not know the word; at the time he was writing his book,
the translation of Greek authors into Araoic had already been going on for
a hundred and fifty years . Ir is because he is avoiding it . For several rea-
sons. First, because in his rime the word musiqi denoted the rules or the art
1~ .! of music bur not music itself as a product of that art, or, if one ma y say so,
l .
as a concrete thing. In order to refer to what we would call music in the
concrete sense, Ghazzali talks either of song (ghina ) or of instruments used
for entertainment (malahi), depending on whether he has vocal music or
instrumental music in mind. And so of course , he had many ocher words at
his disposal, such as "entertainment" (lahw ), in the sense of "entertain -
ment music ,"" "melody" (/ahn ), or "sound" (sawt ), for exam ple, without
mentioning the names of the various musical instruments. Bur he does nor
have a word that covers a very general concept comparable to that we con -
vey by the use of our word "music." (Although, as we all know, this word
Music and Trance among the Arabs 257
can have different, and even opposing meanings on occas ions, acco rdin g
to w ho is using it. )
If Ghazzali does not use the word musiqi it is also bec ause this word, re-
ferring as it does to the rules for composing music , rel ate s much more to
the music maker than to the listener. And it is precisel y the listener he is in-
terested in, not the musici an - for a very good reason , indeed , since in the
eyes of the faith the musician is always a suspect, if not blameworthy per-
so n. His third and last re ason for not using mus iqi is chat it does not con -
note any mor al va lue. The rules of music apply to all kinds of music; not
on ly to chat pla yed for spiritual purposes but also to that pla yed at "gath -
erings w here wine is drun k" (Mole 1963; 174 ) in which the songs are mere
"amusement and futilit y." From Ghazzal i's point of view, w hich is essei-i-
ti ally that of finding a moral justification for samii ', ic is indispensible to
make a distinction bet we en what we might term "light" music and "ser i-
ous" music. Onl y the latter is lawful. To confuse che cwo by us ing the same
term to co ver both would thus be aberrant. T hat which is lawful consisted
of; first, the cantilation (taghb ir)' 3 of the Koran, of course; seco nd, sung
poetr y, on the condition that its sentiments and thoughts were suff icientl y
elevat ed; and third, accompanied song, pro vided that the musical in stru -
ment s utilized wer e permitted, wh ich is to say instruments chat we re never
associa ted wi th blameworth y music al practices. But chis is not all. An -
other restriction mus e be add ed to thes e. It is perm issible to hear only chat
wh ich one hear s w hen one is oneself in a certain state of inner purit y. It is
not only what is su ng or pla yed char counts; it is also the disposition of the
listener . Heard w ith a pure he art , music can be lawful even though it
would nor be if o ne listened to it in a lasc ivious state of mind . This delimi -
tation of the repertoire (Koran , poetr y, ac comp anied song ), m ade all the
stri cter by a pro viso applying to the listener 's own intention, is precisel y
w hat is conve yed by the word samii ' . Th is conceptual p atterning of realit y
to which it corresponds is comparable to no o cher , and cert ainly not to
chat of the word "music. "
Thus , in a Sufi text, the r ange of things possible for "a udition" (samii ") is
restri cted to poetry, the Koran, and, let us sa y, serious music, and these
three things combine, for this ver y reaso n, to constitute a particular cate -
gory of sound. Thi s categor y, which is co vered by no particular Arabic
word (unless it is samii ') , may thus be defined, in th e context of Sufism, as
the category of the object s of audition (samii ') . Thi s catego ry of sounds, as
we have seen, is not only co nstituted by the diffe rent intrins ic qualities of
its three components; the fact chat it is heard also pl ays a role. Jc exists only
insofar as it is percei ved by the ear and insofa r as it affects the hearing.
Twice , on the first p age of his book , Ghazzali stresses the imp ort ance of
the ear : "There is no way of extracting the ir hidden things save by the flint
258 CHAPTER SEVEN

and steel of listening to music and singing, and there is no entry into the
hean save by the antechambe r of the ears" (MacDonald 1901 ). The cate-
gory is, therefore, not constituted at the level of the message itself, nor at
that of its sender, but at the level of its recei ver, or, if one prefers, at the lev-
el of perception.
Why is this? Apparentl y because the three components of this category
of sound are all capable, from the listener's standpoint, of the same ef-
fects. This is what emerges from a reading of the "Book of the Right Us-
ages of Music and Trance, " which clearly shows that the Koran, poetry,
and serious music are all three equally capable of inducing wajd, which is
to say trance, and of doing so in an y context, not only during samli ' (the
ceremony ) bur also outside it. However, let us observe that the manifesta-
tions of this trance will be very different depending on whethe r the "a udi-
tion " takes place during samli ' or outside it. This is a point of great impor -
tance for our purposes, and we shall return to it.

The Word Wajd

This word, which is more often translated as extase or ecstasy, but which I
shall render as "trance" for reasons I shall explain later, is only used in the
Sufi vocabulary and is derived from the root w.j .d., "to find, to meet
with."'" Speaking about the "reality of trance" (wajd ) for Sufis and the re-
lation of "audition" to the soul, Ghazzali wri tes: "Wajd is an expression of
what is found [yujadu] through audition [samli ' ]" (ibid., 719 ). This dis-
cover y, this encounter, involves something that is in close connection with
the deepest part of one's being, or that is in intimate harmony w ith the sit-
uation being experienced . It manifests itself suddenly, like an illumination
(I imagine th at Saint Paul on his way to Damascus was experiencing
wajd ). It is revelation or inspiration. Its relation to "audition," Ghazzali
says, derives from the fact that it "produces purit y of heart and the purit y
is a cause of revealing" (ibid., 723 ). T ranee (wajd ) consists of "Revelation
proceeding from the Truth ," he says elsewhere (ibid., 720 ), 15 and the man
who experiences trance achieves it "because he has found [wajada] wha t
was lacking with him. "
How does this trance manifest itself? Let us begin with the case in which
it is induced by audition of the Koran. "The stories indicating that trance
[wajd] has shown itself in the possessors of heart at hearing the Koran are
many," Ghazzali tells us (ibid. , 733 ), and then cites several of them, begin -
ning with three anecdotes concerning the Prophet himself. The latter is de -
picted as follows: "his eyes were flowing with tears ... ; then he fell, faint-
ing," and weeping, while listening to three different passages from the
Koran. "And much is transmitted from the Companions and the Follower
concerning ecstasy through the Qur ' an. Of them were some who fell
Music and Tr ance among the Arabs 259
swoon ing, and some w ho we pt , and some who fainted, and some who
died in their faint ing. " He then recounts: "And 'U mar heard a man recit-
ing, Verily the punishment of thy Lord surely descends; there is none to
keep it back ! [Qur., 52 : 8). Then he cried w ith a great cry and fell fa intin g,
and was carried into hi s house and ceased not ro be sick in his house for a
m o nth. And $al il} al-Marr I recited ro Abu Jarir, and he sobbed and died.
And ash -Shafi'i heard one reciting, This shall be a day whe n they shall not
speak and shall not be permitted to excuse themselves [Qur., 77: 35); then
he fainte d. And cA]i b. al -Fuc;layl he ard one reciting, A day when mankind
shall rise up for the L ord of the worlds ! [Qur. , 83:6 ] and he fell fainting
.. . . " "And similar stories are transmitted from a number of them,"
Ghazzali adds (ibid., 735 ).
"So, too, is the case w ith the $ufi s," Ghazz al i tells us: "One night of
Ram ac;lan as h-Shibli w as in his mosque, and he was pra ying behind an
Ima m of his, and the Imam ricited , And, ver ily, if We w illed We would
bring to thee him whom We inspired [Qur., 17:88], and ash -Shibli
sh rieked a great shriek , the people thought that his soul had fled; his face
grew red, and his shoulder muscles quivered, and kept say ing, 'With such
wo rds He addresses the belo ved, ' repeating that over and over " (ibid.,
735 ).
"Audition" of the Koran does not have the power of inducing trance
solely among de votees. The follo w ing stor y concerns a slave girl, a tar ' 6
pla yer, w ho we re numerous among the Arabs of that time. This srory does
n ot come from Ghazzali, of course, but from a quite different source. " As
sh e passed close ro a Koran reader w ho was chancing the verse, "Trul y hell
surrounds the impious, " the girl threw away her tar, uttered a cry, and
fainted. When she came ro , she broke her instrument inro pieces and em-
braced religious life.
T ranee induced by the "a udition" of poetry manifests itself in the same
way : by cries an d fa intin g; occasionally, coo , by madness and self-destruc-
tion. For instance, Ghazzali recounts a sror y of how the famous Abu-1-
H usayn an -Nuri, he aring at a gathering a cert ain line from a poem on the
sub ject of religious love, "arose and constrained himself to an ecstas y, and
ran w ildl y on, and happened upon a cane -br a ke w hich h ad been cut, but
the stem s in w hich rem aine d like swords . Th en he kept running in it and
re peat ing the verse until the morning , and the bloo d flowed from his legs
so that his feet and shanks swelled. And he lived afte r that a few d ays and
d ied" (ibid. , 716 - 17 ).
Now we come to an example of trance brought on by singing (ibid. ,
708 ). Th e scene is sec in the vicinit y o f Bag hd ad, beside the Tigri s. On the
verandah of a fine re sidence there is a man ; before him a slave girl is sing -
ing. Bene ath the ver and ah a very handsome yo ung man is listening. He
asks the slave girl to repeat the line she has just sung: " 'Ever y day thou
260 CHAPTER SEVEN

changest!' Then she repe ated it, and the youth kept saying 'This, by Allah,
is my changing in my state with Truth! ' and he sobbed a sob and died."
One last example is also very representative of the situation in which '
trance results from the shock of hearing sung words that unexpectedly
correspond precisel y to the dramatic situation in which the hearer finds
himself. The co incidence triggers an inner uphea va l so intense it can be fa-
t al. Thi s stor y, taken from Isfa hani's Book of Songs ' 8 and wholly profane
in ch aracter, is that of a wo man w hose husb and is raking her away against
her w ishe s, to the co untr y where he lives. The w ife, "having heard a singer
reciting some lines by Abu-Katifah, sig!.s convulsively and drops dead."
The lines of poetry we re: "I spend the nig!-it in grief and moan .. . thinking
of m y fellow cou ntr ymen, w ho live so far away from me ."
All our examples thus far are of unexpected trances linked with situa -
tions in which the Koran , p oetr y, or a song was heard accidentally, as it
were . The audition (samci<)wa s fortuitous . It did not occur during a sama c
in the sense of a "ceremony." When, on the contr ary, trance does occur
during such a ceremon y, it ma y still manifest itself by cries, tears, and
fainting (and eventually even by death ), bur Ghazzali tells us that accord -
ing to the "r ight usages " (adab ) one should not abandon oneself to trance
unless it is re ally too strong; one should always tr y to dominate it. "Then
think not that he who throws himself upon the ground in agitation is more
perfect as to trance than he who is still and does not agita te himself; yet,
often he who is still is more perfect as to tr ance than he who is in agita-
tion ," he writes (ibid., 1902; 6 ).
The second p art of his book is devo ted , first, to the "effects" (athcir ) of
sama c, then to the "right usages" that should govern it. The latter are five
in number (ibid. , 159 ) and relate to (1) the rime, pl ace, and participants in
the sama.<; (2 ) the pre cautions to be tak en rega rdin g the presence of neo-
phytes (murid ) to whom the sama.<could be injuri o us; (3 ) att enti on and
inner concentration; self-control, consistin g in not yielding to the trance
unless it is impossible not to (absence of external manifestations ma y be a
sign of the tr ance 's weakness, but it ma y also be a sign of the subject's
st rength of w ill); (4 ) control of te ars and dancing, neither of which should
be indulged in unles s it is imp oss ible to restr ain them; similarly with the
rearing of garments, which shoul d not be practiced except within certain
limits; (5 ) courtesy towa rd other pa rti cipa nt s and the behavior one
shou ld obs er ve w hile d ancing.
Let us look fo r a m oment at that curious m anifestation of trance called
tamziq, which consists in te aring one's clothes. The same beha vior , which
the evidence suggests is a stereoty pe , can also be observed outside the sam ac
ceremony. One famous examp le is that of a certain Uma yya d caliph (living
in our eighth centur y ) w ho, so the sto ry goes, was so affected by the music
of a gre at singer of the day that he tore his clothing. The singer was given a
Music and T ranee among the Arabs 261
thousan d pieces of gold as a rewa rd, Far mer tells .us (1929; 85), but at
19

the same time the caliph warned him : "Whe n you return to Al-Medina you
ma y be inclined to say, 'I have sun g before the Commander of the Faithful
and so entran ·ced him that he tore his gar ment s,' but, by Allah, if a word
escapes you r lips of what you have seen, yo u w ill lose yo ur head for it ." We
see how much th is form of trance behavior, however conventional and ste-
reotyped it may be, is regarded as irresistible. It is also interesting to note
that it is both and at the same time highl y va lued-even a caliph ma y be
subject to it-and somehow shameful - it must not be disclosed. During a
samZi', one can tear one's garments only if the need is irrepressible, and
only on the condition that the remaining pieces of clothing are still usable
afterwa rds (MacDonald 1902; 10 ).
We have seen that outside the sama ', trance manifests itself quite often
not only by fainting but also by sudden death. Ghazzali (ibid., 4 ) also re-
ports on a case of death during a sama ' caused by struggling too hard to
overcome trance. "It is related," he tells us, "t hat a youth used to accompa -
ny al-Junayd, 20 and whenever he heard aught of the mention (dhikr )2 ' of
God he would cry out. Then al-Junayd said to him, 'If you do that another
time, you shall not accompany me.' And thereafter he kept putting pres -
sure upon himself until from every hair of him there would drip a drop of
water, and he did not cry out. And it is related that he choked one day
through the force of the pressure upon him and sobbed a single sob, and
his heart broke and he died."
Whether trance takes place dur ing or outside of the samZi' , we see it can
produce the same manifestations: sudden death, fainting, cries, tears,
tearing of garments. The only difference is th at "r ight usages" of the sama '
consist in not succumbing to trance but precisely in controlling thes e agita-
tions and manifestations : "Think not chat he who throws himself upon the
ground in agitation is more perfect as to trance than he who is still and
does not agitate himself; yet, often he who is still is more perfect as to
trance than he who is in agitation." Bur after hav ing stared in the first lines
of his first chapter (ibid ., 1901; 200 ) chat the fruit of the sama " is trance,
Ghazzali adds : "and trance bears as fruit a moving of the extremities of the
body, either with a motion that is not measu red and is called ag itation or
with a measured motion which is called clapping of the hands and
swaying of the members." In other words, he establishes an oppos ition be-
tween "agitat ion" (idtirab )- or "non measured" motion - and "measure d"
motions-or dance - to which trance leads when it is controlled . "Agita-
tion," with its ensuing cries, tears, fainting, and madness, is observed
above all, as we have seen, when trance occurs outside the sama ' or, in
other words, when there is no ritual. Let us say that, from our perspective,
this is a nonritualized crisis or trance. Dancing, on the contrary, is ob -
served only during the sama ', which is to say during a ritu al. Lee us say
262 CHAPTER SEVEN

that this is a ritual trance . So once again we find the opposition between
ritual and nonritual trance, wh ich is familiar by now and which we ha ve
encountered so often in matters of possession . Nonritualized trance is fre -·
quently lethal (as Zempleni tells us ), even with in the samii' ritu al, when it
is so violent that it cannot conform to the ritual (as with the young follow -
er of A1-Junayd ). Ritual trance, on the contrary, consists in a state of pleni -
tude and exaltation.
The kind of trance, Ghazzali tells us, th at is "found with Hearing (samii')"
(ibid ., 719 ), and which is a "Revelation proceeding from the truth" (ibid.,
720 ), is at the same time "W itne ssing of the Watcher" (i.e., knowledge of
God ), pre sence of understanding, beholding of the unseen (ibid., 723 ),
communion with the secret, and rel ation to what is lacking. It is pleasure
(ibid., 230; the word appears thirteen times on thi s very page ) and purity
of heart - love, or rather passion, for God. It is not, of course, a posses -
sion trance, and Ghazzali (ibid ., 178 ) who clearl y has Halladj in mind, 22
denounces the heres y of the man who claims to be "inhabited" (hulul ) by
God and to become "one " with him. It is a trance , howe ver, in the sense in
which we defined our terms at the beginning of this book, and not an ec-
stasy, since it manifests itself by mo vement ("it bears an excitation of the
bodil y extremities like a fruit" ), is produced by sensorial stimuli (sounds ),
and is achieved not in solitude but in the company of other participants
(the samii ' has three essential ingredient s accord ing to Al-Juna yd [ibid.,
1902; 1]: time, place, and company ).
Perhaps, in order to distinguish it from possession trance, we ma y call i1
a tr ance of communion or a "communial" tr ance . It is not conceived as th<
effect of God occupying a person - which would be totally impious - bu
as the result of a more or less immedi ate rel ation to God, one that can hav,
the shattering nature of a revelation, the calm nature of contemplation, o
even, at the very limit, the nature of a union. T he wo rd "commun ial'
seems to me capable of co vering all thr ee.
Although it is a trance , then, and not an ecstasy, one might say tha
when wajd is completel y controlled , mastered, sublimated, it is neverrhe
less at the very frontier of ecstasy. This frontier is crossed w hen wajd bt
comes fanii ', which lies somewhere beyond and is a state of annihilatio
(Mole 1963; 269 ), or "disappearance" of human qualities in God. Th i
sta te of ecst asy is frequently the result of another practice known as dhik
which is often closely associated with samii ' .

Th e Word Dhikr
In a very general way dhik r-a word often translated as "recollection"-
ma y be defined as an exercise of piet y consisting in repeating the divir
name in order to recollect God and at the same time to m ake him recolle
Music and Trance among the Arabs 263
one's existence, in the hopes of attracting hi s blessing. The word dhikr in
fact derives from dhakara, "to recollect, remember." Th e practice of
dhikr, like that of samii ', has been the subject of many works by Arab au-
thors, particularly by the Sufis. 23 To be sure , our interest in this practic is
limited to its use as a means of attaining the trance state.
If we broadly schematize" this very complex question, we can say that
dhikr h1s two principal aspects: solitary dhikr and collective dhikr (Gar-
der 1952; 649 ). The first is also called the dhikr of the privileged and the
second the dhikr of the commoners. The first, the solitar y, that of the
privileged, displays three different degrees of ascetic discipline and ulti -
mately leads to what we have agreed to call ecstasy. The second, the collec -
tive, that of the commoners, is on the contrary an attempt to reach the
trance state and contains only one degree of asce tic discipline. Solitary
dhikr, which uses a very elaborate technique of breath control" and wh ich
invol ves the silent and inner repetition of the divine name, leads to a state
of annihilation ,/ anii ', which consists in a total absorption of the self into
God (ibid . , 1953; 205ff. ). Attained in solitude, silence, and immobility,
this ecstas y-or as some say, following Mircea Eliade ( 1948; 93 ), this "en -
stas y"- is accompanied by hallucinations that are mainl y audiror y when
the dhikr is of the second degree, and visual whe n it reaches the third, the
one called "inner" or "secret " dh ikr. Clearl y chis form of dhikr is of no in-
terest to us here, since it is practiced in silence .
The "collecti ve" dhikr or dhikr "of the commoners " (dhikr al-
'aw Zimm) (Garder 1952; 650 ), on the contrar y, is of utmost interest to us
since it is closel y associated with music and dance. Thi s is the dh ikr (or
zikr )' 6 o ften also called "public" (in opposition to the other, which is "se-
cret "), since it is pr acticed by a number of der vish brotherh ood s from In-
dia ro Morocco. Its very spectacul ar aspects, with its violent trances dur -
ing w hich the der vishes pierce their flesh, walk on burning coals , grasp
red -hot pieces of iron without burning themsel ves, swallow broken
glas s- in short give visible proof of their in vulnerabilit y- have been de-
scribed innumerable times . These pr actices were introduced inro dhikr at
a rel atively late date, sometime around our twelfth centur y, it seems . The
style and repertoire of the singing , the use of musical instruments, the
d ance techniques, the demonstrations of , let us say, fakirism ," var y from
brotherhood to brotherhood , each ha ving developed its own particular
form of dh ikr over time. But among those that practice collecti ve dhikr
aloud, the manner of intoning the di vine name by shouting, if not howl-
ing , it o ut remains more or less the same everywhere. Hence the name of
"ho wling der vishes" that has sometimes been applied (very improperly, I
might add ) to the members of these brotherhoods.
-Tothing could be more different from the samii ' described by Ghazzal i
ver y earl y in the twelfth century, or from th e samii ' of the Mawlawiyya 28
264 CHAPTER SEVEN

(the "whirling dervishes" ) as it was instituted some hundred and fifty years
later by Jalal al-Din Rumi, then the zikr of the Rifa ' iyya or the f?acjra29 of
the ' Isawiyya, as the y can be observed in our day . We will now turn to the
relations between music and the di verse m an ifestations of trance that
characterize the se two broad types of ritual, sama c and dhikr . But before
we do so, we must first explain in greater depth rhe opposition we posit
between sama c and dhikr.
According to Massignon, in the course of time the practice of dhikr
gradually replaced the sama c among the Sufis. "Th e masters of mysti-
cism," he writes (1934; Tarika ), "gradually abandon ed free musical se-
ances [sama<] . .. for fixed recitations of litanies based on the Koran
[dhikr]." We shall return to this point before long, bur lee us specify first of
all that sama c and dhikr are as different from one another as the wo rds
used to designate them, and second, that the y have coexisted in the past
and still coexist today. As Mole scares ( 1963; 150 ), dhikr an d sama c "mus e
be carefully distinguished, even though a certain correlation does exist be-
tween chem, and despite rhe face char anti-Sufi polemicists frequently con -
found sama c and public dhikr." Ghazzali, in his tre atise on samac, refers
on seve ral occasions to dhikr. His brother Ahmad also mentions dhikr in
his short description of a samci.' seance (Robson 1938; 105 ). In both of
these cases dhikr seems to be an episode within the general framework of
the samci.'. Thi s was some nine centuries ago. More recentl y, Brunel
(1926; 116 ) found that in Morocco the dhikr, accom panied by very vio -
lent trances, was practiced by the ' lsawi yya, whereas the sama ' , with very
calm trances, was practiced by the Ti ia:niyya. In our day, in Ir aq, the word
dhikr is the one used to denote the tr ance rituals practiced by the Qadir -
iyya and the Rifa ' iyya, while sama c is used for chose of the Yezidi 30 (Has -
san 1975; 221ff. and 250 - 51 ), about which we admittedly have little in-
formation, bur enough co presume char they are ver y different from the
first.
Ir is clearly by the extension of its orig inal meaning chat the wo rd dhikr
has come to denote nor merely the "mention of rhe divine name," as Gar-
der purs it, bur also the entire ritual during which chis recitation occurs,
when it is collective and performed aloud. In the West, and also in some
Muslim regions of the East and of Africa, dhikr has become known as a
ritual char is more or less open to the public and char gives rise to spectacu -
lar manifestations, and it is in this sense, consequently, chat the word is
most often understood outside the narrow -::ircl e of Sufism.
This being so, there nevertheless has been, and still is, a grea t deal of
uncertainty in the use of the rwo words. For example, dhikr is often used
nowadays when referring to the sama c of the Mevlevi. Judging from the
extracts quoted by Mole ( 1963; 232ff. ), the ancient wri ters spoke only of
sama ' , so chat rhe house where the dervishes mer was called sama <khana
Music and Trance among the Arabs 265
and the principal dancer samii. ' zan (he who make s the samii.' ). Inversel y, in
the fourteenth century, Ibn Batutah, in the sto ry of his encounter w ith the
Rifa ' iyya in India, gives the name samii.' to the displa ys of fakirism cus-
tomary among this sect, whereas today we would speak of dhikr, or, bet-
ter yet, f?acjra. This last word, which me ans "p re sence" (not that of God,
as one might think, but that of the Prophet ), and which in man y brother -
hoods designates the week ly dhikr seance, has, according to Trimingham,
"taken the place of the term samii. ' of older usage" ( 1971; 204 ).
It is not only with respect to terminology, howe ver , that things have
changed over the centuries. As Massignon says, there have been even
greater ch ange s in the ritual practices themsel ves. The practice of samii.' ,
as described by Ahmad (brother of the great Ghazzali ), which rook the
form of a "spiritual concert" that gave rise ro highl y controlled mystic al
emotions closer to ecstasy th an ro rrance ,3' progre ssively lost ground to
collective dhikr, which became an extremely violent affair that generated
rather frenzied trances. This change, which began in the twelfth century,
consisted, T rimingham says, in "t he mechanization (if one can put it that
way ) of m ystical experience" ( 1971; 199 ), which from then on became ac-
cessible to "the ord in ary man in a relatively short space of time , by rh yth -
mical exercises involving posture s, control of breath , coordinated mo ve-
ments, and oral repetitions. " It is worth observing, in passing, that this
evolution paralleled a similar trend in the domain of possession, in which,
it has often been said, violent trances are signs of a relati vely late stage in
the cult's development.

Sama.< Music acco rdin g to Ghazzali

In writing his "Boo k of the Right Usages of Audition [samii.'] and T ranee
[wajd]," Ghazzali had in mind readers who already knew perfectly well in
what a samii. ' consisted, so that he did not bother to provide a full descrip-
tion of such a ceremony. Howe ver, the all usions he makes to the way in
whic h it was conducted are numerous enough to give us some idea of ir.
His you nger brother, Ahmad, on the other hand, who also wrote an apol -
ogy for rhe samii. C, has provided us with a very succinct scheme (Robson
1938; 105 - 13 ). Inform ation can also be gleaned from other texrs 32 rhar
deal wi th rhe samii <, either ro defend it or, on the contra ry, ro condemn it.
We thus have enough data to be able to say that at rhe end of the eleventh
centur y a samii. ' was a kind of "spiritual concert," as Mokri (1961; 1014 )
terms it, in which rhe music ,,·as mainly sung, sometimes by a soloist,
sometimes by a chorus, bur which also included an instrumental element
of va rying importance. The concert took place under rhe direction of a
master (shaikh ) who led the ceremony and was at the same rime the spiri -
tual director of the faithful w ho attended. The solo singing was provided
 266 I CHAPTER SEVEl\

by a cantor, the qawwal, chosen for his musical talent and beautiful voice .
The concert consisted of several successive phases , some vocal, some in-
strumental, some a combination of the two. The faithful listened to the
music seated, in a state of great inner contemplation, and allowed them
selves to be gradually overcome by trance (wajd ), which they did their bes
co control. When the trance became coo intense , they rose (MacDonalc
1902; 9 ) and began to dance. Return co calm and normality was brough
about by the sound of music suitable for th at purpose (Robson 1938
112 ), after whi ch ever yone went home filled wi th a memory of "the revel a
tion of what appeared to chem in the state of their absorpt ion in trance '
(ibid. , 113 ).
Besides the singing, what musical instruments we re used? Ghazzali an,
his brother Ahmad are more or less in agreement on chis point. Ahma,
(Majd al-Din ) says chat all instruments (malahi) are "fo rbidden by com
mon consent" with the except ion of the tambourine (duff), whose "perm is
sibility" is based upon two solid references to tradition (hadith ) (ibid.
111, 175 ), and the trans verse flute (ga~ab ), which he calls the "Persia
flute" (ibid ., 96 ). "As for the reed-pipe (mizmar ), it is forbidden to listen t,
it on account of what has come down in the tradition, that He (Allah bles
him, etc. ) heard the sound of mizm ar an d stopped his ears" (ibid. ). Gha2
zali also proscribes wind and stringed instruments (MacDonald 1902
237 ) as attributes of"people who drink and of Mukhannaths [professior
al musicians ]." He also forbids the hourglass drum (kuba ) on the sam
grounds, bur confirms the lawfulness of the tambourine, either with jir
gles or without. He also lists as permitted the small flute (shahin ) accorr
panied by the kettledrum (tab! ), instruments that w ere used, in conjunc
tion, to accompany pilgrim chants (ibid. , 22 0-21 ) and consequentl y lie
because of their connecti on with religious practices.
This exclusion of all musical instrument s apart from the drum and tr.
flute was what Ghazzali and his brother (and others coo) advocated i
their concern to protect the sama c from the accusa tions leveled against
by certain very hardline Muslims. Despite what the y say, however, the sam ,
as spiritual concert was not always restricted solely to the drum and flue,
We know that ocher instruments were used from the wo rk of one of ti-
great historians of Arab music, al-Mas ' udi,;; who wrote about fifty yea:
before Ghazzali. He tells us that when asked what he considered to bed
best form of sama ' from a musical point of view, a governor of the ' Abb:
sid caliph of the time replied : "The best music is that of the four -srringc
iI [lure ] when it accompanies a good song rendered by a perfect voi ce
(quoted in Farmer 1929; 140 ). We are thus informed on the canons,
beauty with respect to sama <during the "Go lden Age" of the ' Abbasids.
Three hundred yea rs later, Jalal al-Din Rumi, the famous Persian po
and founder of the brotherhood of whirling dervishes, would celebrate

____ 1_
 . '

Music and Trance among the Arabs 267

his poems on the samZi' the rebab - a stringed instrument -t he flute, an d
other instruments as well; 34 we shall return to chis point late r. Thus, with
regard to the mu sical instruments used in the samZi' , practice var ied a
great deal over the years. Dou btle ssly it also va ried a gre at deal with re -
spect to place and circum st ance .
Let us return to Ghazzali, howe ver, and stick to his views for the mo-
ment: for him , the music for a samZi' is essentially vocal, and can be ac-
companied only by a drum, and eventuall y a flute . Thi s being so, ho w did
Ghazzal i conceive the rel ation of this mu sic to wajd in the framework of
the samZi' ? What role did mu sic pla y for him in the arous al of trance? To
answer these ques tions, we need to consider his ideas on the effects of mu-
sic as a whole, as the y are expressed in var ious passages of the "Boo k of
Right Usages."
Altho ugh of little importance at first glance, since it oc curs only inciden-
tally in his discourse, there is one obse rvation by Ghazzali that pro vides
the key to his entire theor y. He makes it in the course of his long account
on why so ng, or rather sung poetr y, is more suited than the Koran for in-
ducing wajd. There are seven rea sons for chis , he says (MacDonald 1902 ,
738 ), and the "fifth is that the measured melodies are helped and strength -
ened by the rh yt hms" of the drum (ibid. , 742 ), something th at can ne ver
hap p en w ith the recit atio n of the Koran since it can never, as we know, be
accompa nied by an instrument of an y kind. He then ad ds-and this is the
important point for us - th at "even a weak trance is not aroused ex cept by
a powerful cause, an d it only becomes strong by a combination of these
causes [whic h he h as enumerated ], and each one has a share in the arous -
ing" (ibid., 743 ). For Ghazzali, then , music does not exert its pow er , one
might say, ar any one point ; its action is not single but multiple; it cre ates
not one effect but a cluster of effects. Ir acts at several levels and by virtu e
of several qualities. H ow? First , at the physiological level·, through the
pleasure caused by its sound as an acoust ic phenomenon , when it is beau -
tiful. T his effect is a "secret that belongs to God, Most Hi gh," Ghazzali
cells us (ibid . , 218 ). "Some sounds [aswZit] make to rejoice and some to
grieve, some to put to sleep and some to make laugn , some exc ite [itarZib]31
and some bring fro m the members m ovements according to the measure,
with the hand and the foot and the he ad" (ibid. ). This effect, he goes on to
say, is independent of the meaning of the words, since, for one thing, pure-
ly instrumental music (he specifies the lute or 'ud ) is capable of creating the
same effects, and, for ano the r, the child in its cradle , w ho has no under-
standing, is calmed by pleasant sounds, and even camels, in sp ite of their
"stupidit y of nature," are so affected by the sing ing of their camele er that
they forge t th e weig ht of their burden and th e length of the journ ey; in -
deed, excited in thi s way, stretchin g out their neck s and w ith ears onl y for
the singer, the y can "k ill t hems elves from the force of the pace."; 6 Yet these

_ ,_
268 CHAPTER SEVEN

songs sung to camels, Ghazzali tells us (ibid., 217 ), are nothing other than
poems provided with pleasing sounds (asw at tayyiha ) and rh yt hmic melo-
dies (alhan mauzuna ). Therefore music, by the action of its sound alone, ·
w ithout reference to its meaning , is capable of wonderful effects. But it is
the understanding (jahm) of what one hears that opens the path to ecstasy.
This understanding is capable of attain ing different degrees, which vary
with the state (ha! ) of the listener (ibid., 705 ). There is understanding, we
might say, with regard to the words th emselves, the meaning relating to
oneself, and the meaning relating to God. Onl y the very highest degree of
intelligence, which is reached only after passing through all the other
stages, enables one to attain total trance or annihilation (Jana<). In addi-
tion to the physical pleasure of the music's sound and to the profound un-
derstanding of the words being sung, there is one further element that
must be added if the music is to arouse trance. The listener, as he hears it ,
mu st be penetrated with love (mahahha ) of God: "It is in him who loves
God and has a passion for Him and longs to meet Him ... and no so und
str ikes upon his ear but he hears it from him and in Him, " that music gives
rise to wajd (ibid., 229 ).
Pleasure (/adhdha ), divine love, and beauty are the three wo rds that re -
cur constantly in Ghazzali's accou nt (ibid., 230 - 33 ) of how "audition"
produces trance, after he has first described how, in ocher contexts, it can
arouse profound love, joy, gr ief, courage, warlike feelings, and how, in
other circumstances, it can also be a summons to pilgrimage. The cause of
these states (ahwal ) that invade the hean when one is hearing music is the
secret of God Most High, Ghazzali tell s us (ibid., 23 0 ) and "co nsists in a
relationship of measured tones to souls" (al naghamat al mauzuna ). And
"knowledge of the cause why souls receive impressions through sounds,"
he adds, "belongs to the most subtle of the sciences of the Revelations
which Sufis are granted" (ibid. ). The pleasure (/adhdha ) given by music is
something chat only madmen , the insensiti ve, and the h ard of heart do not
experience. Such people are amazed that it is possible ro feel pleasure and
go into trance as a result of listening to music, and their amazement is like
that of the impotent man who m arvels at "the ple asure of sexual inter -
course , and the yo uth [w ho ) marvel s at the pleasure of govern ing." Be-
cause "in the case of him whose p ow er of perception is imperfect, that he
should ha ve pleasure through it cannot be imagined" (ibid. ). We clearl y
see just how sensual Ghazzali's theor y on th e relations between music and
tr ance is. Elsewhere , w hen he explains w hat the onset of trance owes to
underst anding of the words as opposed ro the pure sound of che melod y,
Ghazzali still talks about pleasure (ibid., 707 ): one faints "from che force
of joy, pleasure, and gladness," he says. Bue this joy, this pleasure , this
gladness, are associated with love of, or rather passion for, God . Pleasure,
understanding, love of God - these, then, are the three components of
Music and Trance among the Arabs 269
tr ance (wajd), which is a revelation of God w hen it is the result of the "au-
dition " of music, poetr y, or the Koran.
If singing (ghina ') , writes Ghaz za li, has greater power than the Koran to
cause trance, this is due to seven reasons (ibid., 738 -4 8): (1) because the
verses of the Koran do not always match the state of the person hearing
them, so that he ma y not understand them; (2 ) because the Koran is so
well known that there is no surprise effect when one hears it; (3 ) because
the sound (sawt; i.e. , of the vo ice ) is more pleasurable when it is measured
(has a regular meter ) th an when it is not measured, and such "measu re " is
wha t distinguishes poetr y from the Koran; (4 ) bec ause the variety of
"measured poetr y" h as great expressive power, even if the melody is pure-
ly instrument al and con sequentl y has no meaning; (5 ) because "measured
melodies are helped and strengthened by the rh ythms" of the drum,
w hereas the Koran ne ver is; ( 6 ) because song can be used more freely than
the Ko r an; (7 ) bec aus e "the Koran is the word of God " and because "it is
uncreated"; bec a use its composition does not lie in the re alm of language
since it is a miracle, whereas poetry, which is composed by men, is in har-
mon y with their natural qualities.
If both the Koran and sung poetr y can induce trance (wajd ), th is is be-
cause both are cap able, through the medium of a pleasant voi ce (ibid. ,
742 ), of making us meet (w.j .d . ), of revealing to us , that wh ich is in pro -
found harmony with ourselves; but if sung poetr y is more capable of doing
this than the Koran, this is especiall y because poetr y is measured and h as
rh ythm. With respect to the arousal of trance , rh ythm and measure ar e
thus regarded as an imporrant dimension of mu sic. But it is not more im -
porr ant th an the qualit y of the sound, the beaut y of the voice, and the
me aning of the words. It is the combination (ibid . , 743 ) of several causes
th at gives birth to tr ance , as we have just seen. Th e drum , whose rh ythm
helps the mel ody and str engthens th e measure , is merel y one among man y
othe rs, an inter est ing point w e need to remember. Although for Gh az zali
th e drum is the only instrum ent , excep t for the small flute, that could be
played in the samii<, it ne verrhele ss only occup ies a so mew hat second ary
place in hi s list o f the effects of music and , ultim atel y, in his theory of the
relations bet we en music and trance. Thi s is a furt her exa mple of the inan -
ity of the theor y, alread y mentioned several times in this bo ok , th at the
drum is always and eve ryw here the principal, if not the o nly, cause of the
arous al of trance.

Whereas in the samil ' , as it is conceived by Gh azzali, mu sic is principally

vo cal, in the sam il ' as it was conceived by Jalal al-Din Rumi one hundred
and fift y year s later, the music is at leas t as much instrument al as it is vo -
cal, if not more so. In the samil ' of the Me vlevi, the famous whirling der -
vishe s of Kon ya in Turke y, the music is in fact pro vided by an instrumental
270 CHAPTER SEVEN

ensemble generally comprising at least two trans verse flutes (ney ), a

bowed string instrument (rabiib ), a plucked string instrument (tanbur ),
two pairs of small kettledrums (kudum ), and cymbals (zim ). 37 It is the ·
flute, however, that plays the principal role and that characterizes the mu-
sic of this brotherhood to the highest degree. "Listen to the flute .... It
complains of the separation," says Rumi in the first line of his great mystic
poem , the Mathnawi. And elsewhere, speaking of the rabiib he says: "It is
onl y dr y string, dry wood, and dr y skin, but from it speaks the voice of the
Beloved. " 38 All of the second part of the Mevlevi's sama c, that which gives
rise to the whirling dance, is accompanied by instrumental music combined
with a chorus of singers. The drums of co•1rse pla y an important rh ythmic
role , bur they are played moderato, and their presence is always fairl y dis-
creet, wh ich conforms to the general esthetics of rhe ceremon y. Unlike the
seances of so many other Sufi brotherhoods , the sama c of the Mevlevi is in
fact remarkable for its extremel y restrained character, and although the
spinning of the dancers is sometimes rather spellbinding in its visual effect,
the spectacle provided by the dance as a whole is always impregnated with
great calm. Ever ything about ir is extremel y controlled, restr ained, or -
dered, in the image of that great celestial mechanism which, as we know, it
symbolizes by imitating the spinning motion of the planets. In this cosmic
serenit y, although the dancers ma y be in a trance, there are almost no out -
ward signs of it. The trance is so interiorized that one is more inclined to
speak of ecstas y than trance; and this indeed is the word most used in this
instance, w hether in French or English , to translate the word wajd.

The word sama c, as I have said, signifies in its general sense the act of
hearing or listening - "audition." The sama c as ritual consists essentially
in hearing or in listening to music, w ith the aim of attaining that kind of
illumination, of divine cont act, that is termed wajd, or tr ance . To be sure,
this "a udition" is not tot ally passi ve, since it requires attention, inner con-
templ ation , and concentration (ibid., 3- 4 ). Moreo ver, once the trance
state has been attained it is exteriorized by means of dance, w hich is noth-
ing other than a way of translating music into action. Ho weve r, the fact
still rem ains that the person participating in a sama c listens to music but
does not take part in making it either as musician or as musicant, except
perhaps when he claps his hands. Reciproc ally, the musicians-singers or
instrumentalists - are there solely to pro vide the music and do not reall y
p articipate in the sama c in th at the y are nor seeking to attain wajd. The
roles are thus strictl y divided: on the one hand there are der vishes, who lis-
ten to the music and expect to enter into trance because of it; on the other,
rhe musician s, w ho m ake the music and expect nothing from it. In this re-
spect, the situation in the samii ' ceremony is identic al to the one that can
Music and Trance among the Arabs 271
be observed in possession ceremonies, and like the latter it is diametrically
opposed to the situation that characterizes shamanism.

The Practice of Dhikr

Although the practice of collective dhikr undoubtedly varies a good deal
from one brotherhood to another, it nevertheless retains certain constant
features, not only in its general organization but also in the manner of re-
citing the dhikr proper. 39 According to T rimingham , "t he harjra [the dhikr
ceremony as a whole] at its simplest consists of two parts: a) the reading
of the office (hizb, 4 0 wa zifa, ere. ) of the order and other prayers, perhaps
interspersed with music and songs (amashid );" and b) the dhikr proper,
accompanied throughout by music and songs and gene rall y introduced
with a special prayer 42 called 'The Opener'" (1971, 204 - 5). Like the sa-
ma ', the dhikr takes place under the direction of the brotherhood's spiritu-
al leader, the shaikh, assisted by one or several singers 43 (munshid ) who
take over in the sung parts of the ritual.
The first part (hizb or wird ) of the ceremon y constitutes a sort of prep-
aration for the second. A state of contemplation and musical fervor is cre-
ated by means of prayers (which are, in fact, ch anted ) and of chants w hich
are taken up in chorus by the participants seared in a circle on the floor
around the shaikh. A certain tension ar ises from time to time, fugitively, at
moments when Allah's name is chanted, while torsos sway back and
forth, prefiguring the essential feature of the second pan . Thi s introduc-
tion can var y in length according to circumstances , bur it seems that, ex -
cept for the moments of tension just mentioned, it remains musically
slack, if one can put it that wa y, from beginning to end.
The second part, on the contrar y, involves an almost uninterrupted cre-
scendo marked our by moments of paroxysm of ever-increasing intensit y.
Everything is now organized around the rhythmic recitation of the dhikr,
which from then on w ill alternate with sung sections, sometimes solo,
sometimes choral. At a sign from the shaikh, the men-for it seems that
the dhikr is exclusively a male affair" - rise. Still keeping their circular for-
mation, which may be right or lax, depending upon whether the men are
holding each other's hands or shoulders, the participants all join in unison
in a rhythmic intonation either of the first part of the profession of faith
(shahada ); La ilaha ilia Allah ("There is no god bur God" ), or of the name
of Allah, or of such other words as huwa (Him ) or hayy (Living ). The
words are half-chanted, half-shouted to a very pronounced rhythm, which
the men obtain by strongly exaggerating their respirator y movements as
the y breathe in and then our. Produced far back in the throat, the vo ice
used is very raucous. The sound that escapes from their chests as the y rise
272 CHAPTER SEVEN

and fall to the rhythm has been described by Ferchiou as "a collective rasp"
(1972; 55 ). In certain cases, this general and raucous panting is- as the
Arabs express it themselves - very similar to the sound of a saw movi ng ·
back and forth." This effect is produced by means of a specific practice
said to dare back to the twelfth centur y and to have been promoted by Ah-
mad al-Yasavi, the founder of the Yasavi yya order and who came from
Turkestan. In order to produce this "rasping saw dhikr" (al-dhikr al-min-
shari ), we are told in a nineteenth-centur y treatise by the founder of the
Sanusi yya (T rimingham 1971; 197 ), "the ha is expired very deeply, then hf
aspired as low as possible." But there are many different ways of vocaliz -
ing the syllables that serve as the basis for the recitation, and they vary
during the course o( the "spiritual concerr."" 6 T rimingham (ibid., 210 ) de -
scribes the rasp during the final phase of the dhikr as being like the noise of
a "rasping saw" or "pecrnral barking." But perhaps it would be more cor -
rect to speak of "roaring." Whatever the case ma y be, the essential fact is
that this great hubbub has an air of savagery and animality about it. As
the y roar, the participants sway back and forth or from left to right, with
their necks disarticulated and th eir heads violently hurled about by the
movement. It is the shaikh who controls the alternation of song and rhyth-
mic recitation, which follow one another without break and are accompa -
nied by various dance figures (spinning, facing one another, splitting up
into small tight groups, or forming a circle , sometimes stamping on the
spot, sometimes not ), and it is he who also regulates the rhythm and gener-
al acceleration of the movements. "Under his dri ve," Ferchiou (1972; 55 )
w rites in her description of a Saduliyya ceremony in Tunisia, "the exercise
of the dhikr reaches its parox ysm: faces streaming with swear, eyes almost
closed , the faithful are drawn into a rh yt hmic frenz y, and then comes the
silence of their ecstasy."
In other seers, on the contrary, this is the moment when the fakir begin
to display the most frenzied manifestations of tr anc e. These manifesta-
tions do not , in fact, have very much in common with the communion
trance th at is, as we have seen, the aim of the dhikr, and these practices are
the ones we must now examine. As we shall see, these fakirisr practices,
al read y alluded to earlier, constitute a very particular aspect of Sufism. Al-
thou gh they ha ve nothing in common with the dhikr in the strict sense of
the term (repetition of the divine name ), the y do nevertheless form an inte -
gral part of the ritu al commonly designated by th is term, or by the term
l;acjra, which in this case is practicall y synonymous with dhikr . As we
have seen, in the dhikr in the strict sense it is characteristic for the adepts
to be the musicanrs of their own trance. As we shall see, this ceases to be
true, however, of the fakirisr section, if I ma y so put it, of the ritual. Our
problem, there fore, is to find out why this should be so . And the answer is
that in the case of the faki r we are no longer dealing with communion
Music and Tranee among the Arabs 273
trance but with possession trance, which means that, log ically, the rela-
tion of the trance subject to music should no longer be the same . In order
to prove this, however, we need ro examine the evidence in greater detail.

F akirist Practices and Possession

All brotherhoods do not indulge in fakiri st practice s to the same extent. In
Tunisi a , for example, as we know, the Saduli yya brotherhood abstains
fro m them totall y. Among those who do go in for such practices, the
Rifa ciyy a, founded in the twelfth centur y in Iraq, and the cJsawiyya,
founded two centuries later in Morocco,' ; are both famous for the spec -
tacular and am azing nature of their exercises. It is with the first of these
th at w e will begin our examination of the relation of music ro this particu-
lar form of q·ance that one might term fakirisr.
The account given by Christian Poche (1974; discograph y ) of a zikr
cerem o ny recorded among the Rifa ciyya in Aleppo is very significant in
this re spect. The ceremon y (hacfra ) in question included the skewer or-
de al, '" which is a rite of transition, o r initiation, that novices (murid ) are
obliged to undergo . It is this trial that interests us here. The trial consists
in the shaik h piercing the neoph yte's flank with a skewer, so that the point
goes right through the flesh and out the other side , without, however,
lea ving a wound or drawing a drop of blood. The deed takes place in an
ambi ance of noisy effer vescence, and the exaltation of the participants is
maintained by continual singing, cries, invocations, and a ceaseless beat -
ing of drums. The novice is put into a trance, Poche writes, so that he can
"identify himself with Ahmad Rifa ci," the founder of the brotherhood. If,
as we stated in the first part of this book, possession essentially involves
identification through trance with a god, an ancestor, a hero - whomev-
er - then this identification with the founder of the brotherhood conforms
perfectl y with the definition of possession. We must therefore view the
Sufi adept in this case, however unexpected or misplaced this may seem
within the Muslim context, as trul y possessed, and as being not in a com-
munion trance - the supposed aim of the dhikr - but in a possession
tr ance. The fact that Sufi religious discourse, unless I am much mistaken,
rem ains silent on this aspect of thing s does certainl y present a problem,
but it does not change an ything. The facts, if Poche's account is accurate,
are there and speak for themsel ves. We shall return ro this point later.
W h at is interesting from our point o f view is that, corre sponding to this
epi so de o f pos session, there is a particular kind of music that is ver y differ-
ent from that used in the rest o f the hacfra. As Pod-e's record notes clearl y
indic ate , drums pla y the princip al role during this ordeal episode. A sig-
nific ant fact is that these drums (actually frame drums ) are called mi z har,
a wo rd me aning (Poche , ibid. ) "he who makes appear." "The entr y of the
274 CHAPTER SEVEN

rh ythm in double time drives the novice irresistibl y toward the center of
the circle," where he is to undergo the skewer ordeal, Poche tells us. No
mention, as one can see, of dhikr or recitation. In th is summons by the ·
drums, using a rh ythm dictated by convention, we see the usual musical
mechanism of possession at work. I shall return later to the role of the
drums in order to summarize Poche 's interpretation and to advance a
slightl y different one of my own - whi ch of the two is correct is relatively
unimportant for the moment. The important point is that first we are
dealing wi th possession, and second , with a relation to music quite differ-
ent from the one th at characterizes the dhikr, as described in previous
pages.
This skewer ordeal was observed in Iraq, in the fourteenth century,
among the Rifa <iyya , by the well-known Arab tra veler and geographer Ibn
Batutah. 49 But his record of the ceremon y is not limited to this trial. The
author adds that, among the dervishes in trance; "some rook a great ser-
pent and bit at its head with their teeth until the y had severed it." For his
part, P. de Felice (1947; 163-64 ) writes the following while describing a
ceremony he witnessed among the <Isawi yya of Algeria, 50 in about 1910:
"These madmen are first presented with knives and daggers, with which
the y pierce their arms and cheeks, without a single drop of blood flowing
from their wounds. Then the y are offered great scorpions and gre y vipers
on dishes covered w ith sieves. The y throw themselves upon those crea -
tures and devour them. One of them, before consuming his share, ran to
and fro with a snake hanging on to his tongue; a second did the same with
a scorpion." Brunel (1926; 21 ) writes that Ben <Aisa (or Ibn <Isa ), founder
of the brotherhood of the <Aissaoiia (or <Isawi yya ), "had the gift of capti-
vating ferocious beasts and rendering venomous snakes harmless." We are
thus faced once agai n with a case of "fa kirism " that relates to identifica-
tor y possession: it is by identifying himself w ith the founder of his sect that
the <Aissaoiia der vish acquires the power of facing the trial and of endur -
ing the snake bites. There is every reason to believe that this also holds true
for the Rifa <iyya.
But the Rifa <iyya have not only the peculi ar ab ility- I was going to say
the specialty-of undergoing sword and snake bites unharmed : the y can
undergo fire in the same manner. In T rimingh am's wo rd s ( 1971; 86 ), their
ecstatic practices render them "immune to sword and fire ." In his descrip -
tion of the abo ve-mentioned seance he w irnessed amo ng the Rifa <iyya, Ibn
Batutah writes: "Loads of woo d had been set read y and were lit; the y [the
dervishes ] danced into the fire. Some of them rolled in it; others placed it
in their m outh until the fire was extinguished." And he adds: "Such is their
rule , and thi s is how the congregation of the <Al)madiyya 5 ' is distin -
guished." But thi s is all he says on the subject. So how is this "immunity "
concei ved?
Music and Tr ance among the Arabs 275
Thi s rime we h av e n o information tha t would pro vide an a nswe r to the
question. By analogy with the tvvo pr eceding ordeals - th at of the skew er
and rhar of rhe ser pent - which b o th have their origin in a p art icular pow -
er of the pa tr on sa int of the dervishes, one could presume the same ho lds
true fo r the fire ordeal. The saint ma y have shown, in one or a n other cir -
cumstance, his resistance to fir e . Following the same logic as in the rwo
previous cases, his entranced devotees, by ident ifying themselves w ith
him, would give proof of the same power . This, however, is pure hypo th -
esis and must still be ve rifi ed.
I had to linger so me w h at on thi s lase case because amo ng all the deeds of
"faki ri sm," the fire o rde al is probabl y the most frequent and most spectac -
u lar. Ir wo uld indeed be particularly interes ting co know h ow chis resis -
tance to fire is conceived, on w h at so re of "collect ive representation" it is
based . The question can also be ra ised for the An as cena ride s, 52 ade pts of a
Christian sect well known fo r it s practice of the glow ing embers " o rdeal.
During the course of their gre at annua l spring feast, che Anastenar ide s
customar ily pract ice a ritual chat consists in walking barefoo ted ove r a
bed of burni ng coals without burning themsel ves. T he adepts prepare for
this o rdeal by entering into t rance, aided by musical acco mpaniment in
\\·hich drumming play s an essential role. Apparentlv, at lease, chis trance is
a possession trance, since the adept is "seized by the saint" (Kakouri 1965;
22 ), whose icon he carries at a rm 's length above his head, as though better
to identify himself with him . Would che imagina ry representatio n s chat
inspire the Anast en arides who walk o n the coals \\·it houc being burnt be
chose of their identification, conscious or not, with the sa int possessing
chem? (Their insen sibilit y to fire, as a physiological fac t, is an o ther mat -
ter, and o ne chat does nor concern us here. ) The hypothesis is onc e more
tempting. Would the biography of the saint - Constanti n , most oft en , but
not always - include an episode relating to immunit y from fire? This here
again is pure hypothesi s, but it might well deser ve an inquiry. According to
a study of th e Anastenarides re cently carried ou r b y a ream from the
Universi ty of Salonica, the people concerned say rhar the y are mo ved by a
"divine power " (S. Baud-Bo vy 1981, personal communication ). In this
case, che Anas ten ar ide s should be seen not as possessed bur as inspired .
With respect to the der vish es involved in the practice of dhikr, we have
seen th ere was roo m fo r hesitation between possession an d in sp ir ation
tr ance. Let it be repe at ed once m o re , conce rning the subj ect s created
here, char rhe distinctions, h oweve r necessary and useful they ma y be , are
always somewhat blurred. T he Anasrenarides a re but one more exa mple
of ir.
Let u s now return to the Rifa ' iyya and mo ve o n to the music of the f?a -
dra . Ibn Barutah does nor say anyt hing abo ut the mu sic that acco mp a n ied
~he seance he descri b es. He o nl y not es that a t the beginning the der vishes
276 CHAPTER SEVE

"rec ited th e dhikr ." After th at , they "started the sa mii c'' (I said earlier what
we shou ld think of the use , somewhat unexpected here, of the word sa -
mii c) during w hich , as we saw, the y rolled about in glowing embers. We
are thus dealing here with a harjra di vided int o two p arts , the first one de-
vo ted ro the recitation of th e dhikr proper , the seco nd ro faki rist mu sic and
practices. Dermenghem and Barbes, in turn, make the sam e di stinction
\vhen the y describe the harjra of the cAissaoua in Alger ia, a harjra whose
seco nd pan is devoted, precisely , to th e "games" of fire and timber." As re -
ga rds the mus ic (which is treated at length in their description ), it is the
chorus th at plays the principal role during the first pan (19 51; 29 1- 92 ),
and "it is mainl y in th e second part of the ceremony that the inst rum ents
(drums and f1utes ) will play a part." We have just seen char drums play the
princip al ro le among the Aleppo Rifa ciyya during the skewe r ordeal. This
predominance of the instruments du rin g the part of the harjra devo ted to
faki ri sr practices is something \\·e shall also find among the 'lsaiyya of Mo -
rocco, to \\'hom we sha ll nO\\' rnrn .
Before we do m ov e on to Morocco, however, w here rhe evidence of pos -
session is even greater th an in rhe cases we have conside red so far, let me
m a ke it quire cle ar that I hav e no intention of reducing all m an ifest ati ons
of Sufi fak irisr pr actices to possession. No thin g is ever as simple as chat. I
simply w ish to show th at in certain ca ses at least these manifestations can
be interpreted in such a way, and th at, at the same tim e, the relations of rhe
adep t to rhe mu sic follow the log ic of possessi o n tranc e (as outli ned in the
first part of this book ), and not that of communion trance . Ir is this corre -
lat ion that it is important fo r us to isolate.
Among the 'lsa wiyya, the public dhikr is pract iced' • eith er in the course
of rhe weekly Frida y mee tings (harjra ) o r during feasts (lemmiit : "ga ther -
ing" ) "o ffered by private per sons very ofte n on the occas ion of family cere -
monies, " or, lastl y, durin g the gr eat feast (mus em ) held every year to com-
memorate the brotherhood's founde r. In all these cases, the dhikr, w hich
in its broad lines conforms to the description al re ad y given and is therefore
completel y Mu slim , merel y precedes other rirnal ac tions th at, o n the con -
trar y, have not h ing to do w ith Islam. Brunel 's wo rk, in fact, shows that a
large part of rhe 'lsaw iyyas' re ligious prac tic e is based upon bel ief in pos -
session by va rious sp irits, usually those of such animals as the lion, lion -
ess, jackal, boar , camel, and so on. As a neophyte, each ad ept is given the
name of one of rhese, and from rhen on he mu st "im itate the behavior and
m anne rs of th at animal exactly '' during rhe brothe rh oo d's weekly ceremo-
nies (Brunel 192 6; 170 ). The origin of rhese bel iefs should probabl y be
sough t in black Africa. "
"The harjra [the \\·eekly ceremom· just menti o ned ] is div ided int o two
essentially distinct parts ," Brunel rel ls us, '·rhe dhzkr p rope r and an ecstatic
Music and Tr ance among the Arabs 277
dance " (ibid. , 94 ). It is the dhikr th at enables the adep ts to go into trance.
The tr ans iti on from thi s communion tr ance to the one th at succ eeds it
wh ich is manifestly a possession trance, t akes place fairly abrup tly, bu~
without causing any re al break in continuity. It is chara cteri zed above all,
it seems, by the more o r less progressive ("surreptitiou s" Brunel calls it;
ibid., 9 5 ) inter vent io n of musical instruments: drums ( thbol and bendair)
and oboe (ghada ). The se give the signal for an ot her dance, the rbbiini,
which is relatively slow and is somewhat akin to that of the dhikr, since it
essentially involves "a so n of rh ythmic pounding, now on one foo t, no w
on another," then "in a downward bend in g movement, " while "a r aucous
and troubling cry of Allah b reaks away from their chest." And "the re are
no variat ion s on this dance," Brunel obse rves . Thu s we a re in the presence
he re of what I proposed, in an earlier chapte r , we should call an abs tract
dance . 58 T his serves as a kind of "p reparato ry exercise" fo r the next dance,
\\·hich is figurative this time , and d irectly related to possession . The fig-
ures of this ne\\ . dance change with th e T a' ifat, or "groups," to which the
dancers are arrached . Sometimes the dance invo lves two performers w ho
face one anothe r ; in this case, thev execute spinning leaps, then p retend co
·'repel one anothe r reciproc ally like two opposed forces," a cho reographic
figure alluding to the son of Ben 'Aisa, a "famed sailo r " who ·'when the sea
broke upon the sho re, hurled himself daringly coward it," so that "the
mass of powerful water crashed against his bre as t. ,. Sometimes the dance
involves "exrraordinarr leaps' · accompa nying the gestu re "of dis -
embO\\·elling and rearing our,'' which are execu ted by the frriissa : "those
\\·ho de\·our,"' 9 \\·ho are lions (sbua ), panthers (nmoura ), o r jackals
(dhiab ). Other d ance figures corresponding to th e behavior of va rious ani-
mals capable of possessing the adepts -bo ars, jackals, and so on - also
occu r, of course. 6 0 Those \\·ho embody the camel, the jma/, reproduce "th e
movements of the a nim al they are representing in the most detailed man-
ner" (ibid., 204 ). Their specialty is to chew cactus leaves bristling with
spines, o r barley, or wild artichokes . They- also sometimes quarrel among
themselves - "TheY butt one another with their shoulders and roar, bite
one another sa\·agely, and exchange viole nt kicks" - and somet imes "hurl
rhemseh ·es, head down, at the gates and bacrer at them with great vio -
lence " (ibid., 205 ).
There is no point to going any furthe r: clearl y thi s second part of the !Ja -
clra, \\·hich Brunel describes as devoted to "ecstatic dancing," has no con -
nection with the first, \\·hich was devoted co th e "dhikr proper." What we
need to remember, once agai n , is that the relations between trance and
music totallv change from one part to the other. In the dhikr, the mu si-
cants are the adep ts themselves: they are the ones who make th e mu sic
that governs the exercise that causes them co ente r into a trance. In the "ec -
278 CHAPTER SEVEN

static dancing" - which, all things considered, we migh t as well ca ll t he

"possess ion d ance "- it is no longer the adepts who provi de t he mu sic , by .
singing, but the musicians, who do so by playing drums and oboes.
Before lea ving religious tr ance among the Arabs, and passing on to pro -
fane trance, perh aps it wou ld be useful to review the relations betwee n
possession and Islam as they emerge from the examples we have examine d
thus far. In the case of the <Jsawiyya, we are clearly dealing with a phe -
nomenon of reli giou s syncretism. Let me simply say that such things are
by no me ans unique, that no one is surprised when the y occur wi th in a
Christian frame wo rk, and there is thus no reason to be any mo re surp rised
simply becau se the y occur within an Islamic context. In Bahia, those who
t ake part in the candomble have no difficulry in practicing possession an d
at the same time decl aring themselves Roman Catholics. True, n umerous
Christians wax indignant over this fact, but then a numbe r of Musl ims ar e
equally ready to express stern dis appro va l of the Sufi practices I hav e
termed "fakirisr." The two cases are thus symmetrical in this respect .
Ho wev er , although both crsawi yya and adepts of the candomble, each in
their ow n way, provide us with absolutel y clearcut examples of sync re-
tism, both openly combining possession and ado r ation o f God, the situa -
tion is not a lw ays as clear in ot her cases.
When, in the course of a f?acjra,one section is devoted to the practice of
dhikr, which is a recitati on of the di vine name and commu ni on w it h God,
and another is devoted to the orde al of being pierced with a skewer, which
is an initiation rite and identification w ith the founder of the brotherhood,
in the first case we are indi sp utab ly dealing w ith a Muslim act of faith
whe reas in the second something else is involved . But w hat exactly' That
is the question, and it has several possible solut ions .
I have suggested \\·e view the trance during the ordeal episode as a case
of possession. Th ose involved woul d undoubtedly be both surprised an d
indignant at this suggestion . To be sure, t;1ey exper ience this pri vileged re-
lation w ith the bro ther hood's founde r- with the saint - merel y as a par -
ticular aspect of an overall situation, which is adora tion of Allah. The
saint is n o more th an a mediator in this adoration, and this is why there is
nothing shock ing about identifying oneself with him. It is no nethele ss true
that this identification, which im·olves going into trance and temporaril y
taking on the personality of someone who is the object of the gro up 's de -
votion for th e purpose of obtaining his blessing, conforms exactly to the
definition of possess ion proposed at the beginning of thi s book . I shall go
no furthe r than this, ack nO\\·ledging all the while that the ambiguity of
this case somewhat shades the presence of possession, but the latte r none -
theless exists. So that a f?acjrain which dhikr prope r and fak iri st practices
succeed one another, in w h ateve r o rder ,6' should be seen as in so me way
Music and Tr ance among the Arabs 279
juxtaposing communion and pos sess ion trance. T his is what causes the
sing ul a rit y of these ceremon ies. Such a juxtaposition is likely co b e a
source of ambi guity, so that it is n o t always clear w hether it is th e firs t o r
the second type of trance that is at wo r k. Bue the existence of transitional
or a mb iguous stages does not alter the fac t chat tw o types of trance are
clearly involved, or, if one prefers, two ty p es of re ligious bonds wi th the
deity . Th e interesting point for us, lee me repea t , is chat the relations be-
tween the trance su bject and the music va ry according to whethe r the first
or the second type is invol ved. During the dhikr proper, which is to say
during communion tr an ce, the a dep ts are the o ne s who ar e the artisans o f
their own entry into trance, wh ich the y induce by their own singing and
dancirig; during the faki ri st practices, they n o longer do so themselves but
others nave the task of providing the mu sic, and of inducing their trance.
In short, in the fir st case the a dept is a m us icant and in the seco nd he is mu-
sica ced.

Th e Jinn and "Poet ic Furor"

Alongside the possession trance a ssociate d with Islam throu gh the popular
for m s of Sufism we ha ve just examined, the r e is another kind, parallel
w ith Isla m , but also sometimes associated w ith it. Thi s is the p ossession
trance brought a bo ut b y chose ill-defined bein gs- lee us say those
"geni i"- known in Arabic as jinn . •2 A pre -Isla m ic survival, belief in the
jinn has become confused, or combined, w ith a number of o cher more or
less similar beliefs chat are, or ha ve been, cur ren t in the va r ious cou ntries
in which Isla m prevails. The jinn are held responsible for cert a in illnesse s,
whic h they cause by taking posse ssion of people, and the in divid u als
posse ssed in thi s way rend to '.:lrganize themselves into schools or colleg es.
Th is is tr ue, a m o n g other examples , of the female college of the Tidjan-
iyya63 in Tunisia, des cribed by Sophie Ferchiou ( 1972; 57 ff. ): th e "sp irit s"
responsible for the possession and called 'ennes el ohra ("the ocher people" )
answer for the mos t p art to the definition of jinn . 6'
Seen in this way (for there are ocher aspects, some assoc:at ed with m ag -
ic), the jinn, or rather the cults of which the y are the object, scarcely con -
stitute any thing more th a n variations on the general ~heme of possess ion.
And these va ri at ions would probably tell us no mo re about the relation be -
tween music and possession th an what we have already learned in the first
part of this book . I shall therefore not linger ove r them . Moreover, Far m er
(1929; 7 ) tells us th at "the jinn were evidently conjured by means of mu -
sic," which takes us back once m o re to the exorcism problem, w hich we
also discussed in p a rt one - th e m ore re aso n for m ov in g on . Let it simpl y
be understoo d at this point that in the subseque nt general cable of trance
 280 CHAPTER SEVEN

(cf p. 290 ), trance rela tin g to jinn cult s should of course be pl ace d in the
category of possession. 6 '
Belief in the jinn, however, presents two othe r aspects that are corn-
pletel y different and particula rly interesting from our point of view: first,
that of the shaZi'ir majnun , or "poe t- soothsaye r possessed of the jinn"
(ibid., 21 ) who pe r forms the function of prophe t or augu r 66 (in hi s earl y
days Muhammad had the reputation of being one ); and seco nd , that of
the musician, or more correctl y of the composer or poe t , who receives hi~
inspirations from a jinn. Like his master, the cele brated Ishaq al-Mawsil i.
the most famous of the Andalusian ninrh -century mu sicia ns, Zir yab, "be-
lieved that the jinn taught him his songs in the middle of the night. Wher
thus inspired he would call his two favorite singing-girls , Ghazzalan anc
Hind a, and bid them commit to memor y music wh ich had co me to him b~
the se means " (ibid., 130 ). A good ex ample of what the Renaissance wa
later to call, following the Greek s and Romans,furor poeticus.

Possession by the jinn as expressed in popular religious pr actices, th ,

clairvoyance of the augur, and the inspiration of the poet, can be seen a
equivalent to three of the four kinds of mania described by Plato : the tele~
tic , mancic, and poetic. But w h at has happened to the fourth, the one Plat,
·himself placed above all the others, the erotic,6 ' which, to be sure, is assc
ciated w ith love, but with love as a contemplation of beaut y and throug
this of goodness and truth? This o bviously pa rallels commun ion tr ance
the wajd - the paroxysmic expression of the love of God. The four typt
of trances thus far observed amo ng the Arabs thu s correspond to the fot
sorts of mania practiced among the Greeks . A great deal could be sai
about this correspondence, which is certainly not totally fortuitous, an
about this quadripartite division of trance. I will simply observe th at a
th oug h the first three categories of tr ance, those due to the jinn, certa in
owe nothing to the Greeks, the fourth, that wh ich is due to God, parado
ically does owe them some thin g, in part at least . It is generally acceprc
that the birth of Sufism is some\\·hat related to Neoplaton ism. Concempl
tion of a beautiful youth as a means of attain ing trance, a practice espec i;
ly advocated bv Ghazzali's brother Ahmad, 68 might well be a more or le
direct heritage from Plato .
This is the end of this dig r ession, the purpose of which \Vas to show th
as far as trance is concerned, the situation among Arabs of the eleven
I.
centur y and that among rhe ancient Greeks, had certa in similarities, c
l.f sp ite their enormous diffe rences. After this brief mention of the "poetic I
ror" of musicians, which, though it is occasi o ned by gen ii, has no clc
association with religion , let us now mO\·e on to an examination of co
pletely profane trance .

il.-...._ - ·.
Mus ic and T ranee among r~e Ara bs 28 1
PR OFANE T RANC E

U p to this po int , the tr ances w e h av e ex am ined in the samZi', the dhikr , o r

the f?acjrah av e all b ee n re ligio u s . T h ey a re de sign ated either by the wo rd
wajd, a term th a t co m es from t h e sch ola rl y Sufi tr aditi o n and is n o t in very
w ide u sag e , o r b y t he wo rd ha!. A m ong th e Ar a bs, howeve r, tr ance can be
produ ced b y mu sic o utside o f a n y re ligio u s co nte x t. It is th en referred to as
tarab .

The Word Tarab

Nothing can bette r ex p la in what tarab is, in t he sense just in d ica t ed, t h an
the following anecdo te taken from the Kitab al-aghZini ("T he Book o f
Song s" ), a famous collect ion composed by lsfahan i in ou r tenth ce ntur y.
The scene is set in Med ina, rwo ce nt uries earl ier , ur.der the U m ayyad . The
famous singer, Jamila,•• was in the habit of entertaining the best poe ts a nd
music ians of he r t ime in h er ho m e, amo ng chem the gr ea t erot ic po et cu_
mar ibn -Ab i Rabi'a. O ne day, dur ing a conce rt at her house, she b ega n
singing some ero ti c \·e r ses he had w r itte n:
"As Jam Ila sang, all rhose gathered the re ,,·ere seized by tarab [ecstasy]:
they began to clap thei r hands, bear rime on the floo r ,,·irh thei r feet ,
and sway their heads, shouting: '·\Y/eoffer ourselves in sacrifice for thee,
oh Jamila . to protect )·ou from all evil. . How sublime ,·our song
and your words 1" As for the poet 'Umar, he began to shout our: "\X'oe is
me. W/oe is me . . ." He tore his robe from top ro borrom, in a stare
of total unconsciousness . Whe n he came to, he felt ashamed and began
to apologize, saying: "By Allah, I could nor restrain myself. for that
beautiful ,·oice made me lose mY mind ." The other guest answered him:
"Console yourself, the same happened to us all, and ,,·e fainted . Bur
we did nor rear ou r clothing." [Jarg,· 1971, 25- 26)

Clea r h· the external ma n ifestations of tarab, th e profane trance, ar e in

all poi nt s similar to th ose of wajd, th e mys ti c tr ance, when th e la tt er is ex -
pressed by fa intin g, loss of co n scious ne ss, cries, an d rearing one's ga r-
ments (tam:::,iq ) . Beside rhemselYes with emotion, people start to cla p t heir
hands and dance. Indeed, wajd and _tarab somet imes resemble one an:
ocher so closelv t h at t hat _tarab may also used to signify mystic tr anc e.
This is the case in certain a reas of Iraq. A. distinction is then made between
tara b k aif and tarab r11hi, "entertainme nt tr ance" and "spi r itual tranc e"
(Hassan 1975; 305 ).
Just as wajd, as we have seen, may dis p lay d ifferent degrees of inten sity,
so that its manifestations van· from the most abandoned and spectacula r
form of frenzy, inclu di ng even sudden death, to th e most cont rolle d a nd
282 CH APTER SEVEN

interi o rized expres sion, so _tarab can likewise lead to the worst extremes
of madness, even death, or, on the contrar y, be reduced to a pure and sim -
ple musical emotion of which no sign, or almost none, is externall y visi-
ble. ;o owada ys, ho w ever, especiall y in urban settings, trance as an ex -
pr essio n of music al emotion , of farab, is less customary than it was in the
p ast. It still occurs, ne vertheless, main ly in countr y distr icts. In certain
areas around the Euphr ates, for ex ample, we ar e told by Scheherazade Q.
H as san ( 19 7 5 ; 12 5 ), people gather every evening in the villages to sing ,
and it is usua l for the men to go so far as to "w eep with ecstas y." In Iraq, 1

too (ibid. , 32 ), in the sm all folk orchestras , the professional drummers of-
ten br ea k their drum s' 1 on their ow n he ad s "a t the moment of ecstas y"; a
form o f destructi ve behavior that natur ally m ak es one think of tamziq, the
ren din g of one 's clo thin g, w hich w e ha ve encountered before , and which
in Sufi territor y w e found as a m anifest ation o f mystic trance (wajd ), dur -
ing th e spiritu al concert (sam a') .
Tarab and m usi c are so closely assoc iated th at the w ord tarab h as in
fact co me to sign ify mu sic. Thu s in o ur d ay, in Iraq (ibid ., 123 ), mu sical
inst ru ment s are as o ften referred to as a/at il-.tarab, "instruments of
trance," as they are a/at il-mus zqa . Thi s usage is bv no me ans recent , as is
att ested, am o ng ot her exa mples, by the tr ea tise of Ibn Abi 1-Dun ya , the
great ni nth-c entur y mu sic cen so r, w ho already called musical instruments
a/at al _tarab (Robso n 1938; 12 ). And even as ea rly as the time of the first
calip hs, in the seventh centur y, Farmer tells us (19 29 ; 5 1 ), music was also
called ,tarab. The gre at hi sto ri an M as' ud i, in th e tenth centur y, w rote ,
,,·hen wish ing to ex pre ss a cert ai n caliph 's pass ion for music, th at he wa s
dee p ly in lov e with ,tarab (ibid. , 60 ). And ther e is another more recent ex -
ample: a book abo ut mu sic publi shed in Bag hd ad in 1963 and entitled A l-
,tarab ' ind il-'Arab ("Tarab am ong the Ar abs" ), is not at all about tr ance ,
co ntr a ry to w hat one might ex pect , but abo ut singing.
It is high time I po inte d out th at the wo rd _tarab deri ves from the verb
fariba, w hich me ans "to be mov ed , agit ate d" (Kazimirs ki's diction ary). Of
so me o ne "m ov ed by th e mem ory of hi s h om eland and longing to return
the re" (ibid. ), o ne w ould say he is _ta rib, and of "camels hurr ying to re ach
camp agai n "- ex cite d n o doubt by the rider s' songs mentioned by Ghaz-
zali" - one wo uld say the y are ,tarib. Tariba also signifies "to excite , to
wa nt to m ov e ," and hence (Kazimirski ) "to sin g, to make music." Mufrib
and mu .triba, lite ra lly "he " and "she w ho m ov es pe ople," are the words for
m usicia ns , m ale and fem ale . The fam ous Egypti an singer 0mm Kalsum ,
whose re co rd s so ld by the milli ons , was a mufriba, and it wa s quite usu al
fo r th os e listening tO her tO be seized by _tarab . For m utri b and m u,triba,
are , in fact, applied tO m akers of popular mu sic only, riot to makers of
learn ed music, ' 3 and of cour se more rn singer s th an to instrumentalists. ' '
Music and Trance among rhe Arabs 283
On the other hand , mu _trib can also sometimes mean "music ." Th ere is a
tenth -century Arab trea t ise" in w h ich we read that the G re ek w ord mu siqi
denotes what in Arabic is called mufrib . In Ir an, the wo r d mofrebi is ap-
plied to light or popular music (D uring 197 5; 141 - 42 ).
T he fact that it is poss ible, when speak ing abou t music, to u se th e wo rd
farab (wh ich in fact denotes the emotion or tra nce to which it gives rise ),
and when referring to musicians to use mufrib (which means literal ly "h e
who arouses _tarab"), demonstrates clearly enough the close ties that u nite
music and trance among the Arabs. For them, both words can be synony -
mous. We saw earlier, in the case of the sama ', that music was the grea t
mover of m yst ic trance; and now we find that it fulfills the same functio n
for profane trance. In both cases, we should note, it is the expressivi ty of
the singing that is the operative facror, working through the combined ac -
tion of the beauty of the voice and the emotional power of the words . The
instruments have little, or even nothing, to do with it. Ind eed, contrary to
what one might expect from the expression a/at al -tarab, "the instruments
of trance, " wh ich in fact means musical instrument : , as we have just seen,
trance, as a manifestation of mu sical emotion, is so closely assoc iated
with singing that the musical instrument is sometimes regarded as unfa -
vo rable to the inducement of farab . In classical singing, that of the maqa -
mat, "only the great singers can attain joy and ecstasy [.tarab] in the pres -
ence of instruments," Scheherazade Q. Hassan tells us (1975; 125 ). She
also tells us that in the villages near the Euph r ates mentioned just now,
"w here lo ve of singing con5titutes a genuine mystique," the musical instru -
ment is "a thing of shame" : in their singing sessions, rhythmic accompani -
ment is provided "solely by a string of beads ." In other words, _tarab, pro -
fane trance, is no more linked with frenzied drumming than is wajd ,
mystic :ranee. Just as Ghazzali says that the triggering of wajd during the
course of the sama ' is linked with the sp ell of the vo ice and the power of
the words, so is the triggering of _tarab in the course of a concert of popu -
lar music . This is so true that the Iraqi proverb bait ii _tarab ma kharab,
"the house of the farab does not know misfortune," means in plain lan -
guage, "the house where one sings does not know misfortune" (ibid.,
124 ).

THE E. "GE!\"DERING OF TRA CE

The relation of trance to music and dance among the Arabs, as it has
emerged in this chapter, would seem robe somewhat complicated. This
complication is more apparent than real, howe ver, and the data as a
whole can be see n to fit inro a fairly simple pattern if we simply rearrange
them in accordance with, on the one hand, the d ifferent conditions under
284 CH APTER SEV EI

wh ich tr ance oc curs, and, on th e ot her, cert ain mod alit ies of mu sic an
dance. From our viewpo int here, it is possi ble to reduce both of th ese set
of data to a small numbe r of op p ositions . Th is is wha t I now intend to d~
' The conditions under wh ich tran ce occurs may be either pro fa ne or rel
gious, and when they are religio us th ey can be either "co mmun ial" (mear
ing that they co n sist in a relat ion of commu n ion with Allah ) or fak iri :
(meaning that they relate to the fakirist practices I have de scri be d an
which fall into the category of possession ). If they are co mm un ial, t he
can be either ritual ized or nonritua lized. If t hey are ritu alized , they ca n O<
cur w ithin one of th ree framewo rks: "class ical" samii ', th e samii ' of tr
Mevlev i, or the dhikr in the strict sense of the term.
As for the music that accompanies the trance, it can be made either I::
the person who is in trance or prepa ring to enter it, or , on t he cont rary, I:
other persons spec ificall y ch arged with th is task . Plac ing o ur selves in ti-
position of the trance subject, who is thus either the "add resser" or tl-
"ad dressee " of the music, we shall say th at in the first case he is a "mus
cant" and in the second he is "musicated ."
Lastly, dance - and it is t hat of the entranced person that int erests t
here, needless to say - may either precede entry into trance and p repare i
or, on the contrary, follow it and thus ser ve to manifest it. We shall s.;
that it can be either the cause of the tr ance or its effect. It can eventually t
both, one after the other, bur in this case it changes in character, so tha t
is no longer the sa me dance that is involve d .
This being granted, what kind of correlations can we find between tl-
first set of data (condition of trance occurrence ) and the second (modal
ties of music and dance )'
Profane trance, th at which is found outside any religious context,
triggered by hearing music (nonreligious mu sic, of course ), or, more pr
cisely, singing. The circumstances are of little importance, and the trarn
occurs sometimes completely unexpectedl y, sometimes, on the contrar
in answer to a high degree of expectation . Its most characteristic manife
rations are weeping, fainting, and the rending of garments . It can ali
happen that the trance subject, who is beside himself, will begin to clap h
hands and strike the ground rh ythmicall y with his feet in a dance. Wha t
invo lved, then, is a stereotyped (but not ritual ) pattern of behavior th .
expresses in a particular \\·ay the emotion experienced when listening 1
music. We shall call this tr ance "e motion al trance .''
As just described, emotional trance is passive, both from the sta nd poi
of music and that of dance, since it is only wh ile listen ing to music, n,
while making it oneself, that one enters the trance state, sinc e m ost of ti
time this state solely involYes fa inting, weeping, o r rending one's ga
ments, and also since although there ma y be movement - hand -clapp i1
and dancing - this moYement is only a manifestation, which is to say
Music and T ranee among the Arabs 285
sign or co n sequence, and no t a cause of t he tr ance. T he per son in tr ance is
to be rega rded, then, as esse nt ially "m usica ted" and on ly very acc esso rily
as a "musicanc." T his rule is not absolute, however . Emotional tr ance is
somerimes'Observed in musicians while they are pla ying. Scheherazade Q.
H assa n describes Iraqi drummer s br eaki ng the ir earthenware drums on
their heads "at the moment of ecstasy (!arab ). " This would seem to be a
case of emotional trance in a musicanr rather than in a musicated person.
Is it an aberrant case? To answer this we would first have co know wheth -
er these particular musici ans are entranced by their own drumm ing or
whether, on the contrary, the trance is due co the words of the singers they
are acco mp anying . Boch are equally possible. In the second case, despite
appearances, we would be dealing with emotional trance induced in a mu-
sicated person, which falls perfectl y well with in the ca tegory just de-
scribed . The mus ician would, in this instance, experience the odd situa -
tion of being what one might call - no pun intended - a musicared
mus1canr.
ow lee us move on co religi ous trance, beginning with communion
trance. As we have seen, it can occur outside any ritual context, and I cited
a number of examples of people being unexpectedly seized by wajd simply
upon hearing someone singing a verse from the Koran o r some verse of re -
ligious poetry. The manifestations of thi s trance are much more violen t
than chose of the ritu al ized wajd, sin ce they can include sudden death,
fainting , and fits of madness. The difference between the two is the same
as char wh ich distinguishes nonritualized trance from ritualized trance in
th e possession cults we examined in the first part of chis book. The nonri t-
ualized / ritualized opposition can therefore be used once again, within the
framework of Sufism this time, in order co distinguish the kind of trance
that occurs outs ide the samii<, and wh ich can be lethal, from chat which
occurs w ithin the framework of the sama <a nd which, apa rt from very rare
exceptions, is nor fatal. In face, this nonritualized communial trance dis -
pla ys much more the characteristics of a crisis (and I am thinking here of
what we called the "prepossession crisis" in the first part of chis book ) than
of a trance. Ir ne\·er leads to dancing. The relation of the subject in a
trance - or in a crisis - co the music is a p assive one. The subjec t is a musi-
cared person, not a mu icanr.
Ritualized communion trance occurs inside the framework of three dis -
tinct rituals: what I have called the "classical" samci ' , which confo rms wit h
descriptions daring from the rime of the two Ghazz a li brothers; the samci '
of the Me\·levi, which is a later form; and the dhl/...r. In the classical samci C,
the adept is ent ranced (wajd ) by the "a u dition" (samci' ) of music. His rela -
tion to the music is thus once again passive. Bur here, unlike the nonrir -
ualized communion trance, the trance is necessarily, or almost, trans-
formed inro dance, so that ulrimarely dance can be see n as the outstanding
286 CH APT ER SEVEN

sign o f its rituali za tion. Let us once again recall what Ghazza li w rite s at
the very beginning of his "Book of the Right Usages of Auditi on and
Trance " : T he trance [wajd] "be ars as fruit a mo ving of the ex tr emi ti es of ·
the body, either wi th a m oti on that is not me as ured and is called ag ita tio n ,
or w ith a mea sured motion which is called clapping of the hands and
sway ing of th e membe rs ." The op pos ition he establishes here bet w een dis -
ord ered mo vement , wh ich is mere agita tion, and the ordered ("me asured" )
move ment that constitutes dance could nor be mor e clear. One might al-
most think th at in thi s passage Ghazzali had the nonritu al ized and ritual-
ized distincti on in mind. In the last part of his book , in whic h the "right us-
ages" are actu ally pre scribed, he de als exclusivel y w ith dance. He doe s not
say a wo rd abo ut ag it at ion. Th us dance ultimatel y co nstitutes , for Ghaz -
za li himself, the very bloss oming of tr ance. Seen in its relation to music
and d ance , the tr anc e that oc cur s during the sama c thus appea rs as both
the effect of music and the ca use of dance. I shall come back to thi s later,
bur let me say here and now th at this is usually the case in possessio n as
well.
The situation is n ot the same for the sama · of the Mevlev i , howe ver, an d
we shall not turn to it. Contrary to the "classica l" sama <as it is possible to
reconstruct it from the wr iti ngs of Ghazzali and his bro ther Ahmad, the
samii <of the Mevlevi (the '\ \·hirling der vishes" ) is characte ri zed by the fac t
that dance is not th e re sult of tr ance but in fact its cause . In the classical sa-
ma<ir is tr ance that inc ites the p artic ip ants to dance , in the sama <of the
-Mevl e\·i the oppos ite is true. Ju st like the classical sama c, rhe samii <of the
Medevi begins with a section devoted to p rayers and in vo cations, then to
music, which the dervishes listen to seared and motionless. At a given mo-
ment, they rise and walk on ce aro und the room, still to th e so und of mu -
sic, after which, having removed the ir robes, the y beg in to spin. The im -
portant thing to st ress is that they do nor wa it until they are entrance d
before the y begin to sp in, in other words, to dance. Quire the contrary, i1
is the wh irl ing morion of the dance itself that leads th em to ecst asy, or , as •
\\·ould prefer to say, to trance . Thi s tr ance is th erefor e the result o f th(
adepr's own act io n . In relation to hi s entr y into trance the adept is thu s ac
rive. He is the acto r , through his dance, of his own entry into trance. Th,
subject's w ill is in vo lved in a \\·ay th at is not found in the classical sama c, i1
which rhe adep t simply undergoes the effects of the music by going int<
trance in the first place, then by translating it through the m ove ment s o
his dance. Seen from the pe rspective of the onse t of tr ance, the differenc
is cons ider able .

Induced T ranee vs . Conducted T ranee

In orde r to express this important difference, I propose we use the two OF
posed ter ms "induced trance" and "co nducted tr ance. " Induc ed trance wi
Music and T rance among the Arabs 287
signi fy a trance into which the subject is led through an actio n ex ternal to
him self; conducted trance will be chat into whic h the subject leads himself
through his own acti o n; in ocher wo rds , the trance is self-induced .
Clearly the tr ances we ha ve listed up to now - (1 ) profane (o r emotion -
al) trance; and (2 ) communi al tr ance, both (a ) nonricualized an d (b ) ritu-
alized in the classical sama '- we re all induced tr ances . The commun ial
ritualized tr ance in the Mevlevi samci' , on the co ntr ar y, is a conducted
tr ance. Bur, one may say, in the Mevlevi sama ' there is also au dition (sama ' ,
in fac e), not just dancing. The Mevlevi are "dance rs, " it is true, but the y
are "music ated " and nor "musicants ." My repl y would be char the imp or -
tant aspe ct to remember is the facto r char engenders trance, which in this
case is dance and nor music, w hich is nor the case, I repea t , in cla ssical sa-
ma ' . This being so, I ack nowledge tha t thi s case is nor totally devoid of
ambiguity.
Thi s am bigui ty bec o mes even m ore evident when we move on to an -
other case th at is completely unambi guo us and tha t cons titutes, one might
say, an exam ple of purel y conduc ted tr ance. This is the dhikr of the Sadu -
liyya of Tuni sia, as d esc ribed by Sophie Ferchiou. In chis case there are no
musicants other th an the ade pt s themselves. The y are the on es w ho sing
the pra yers, the incantations, and the songs that for~ the first p art of the
ritual that ser ves as a pr epa r ato ry stage for the dhikr proper. As for the ex -
ercise of che dhikr itself , w hich, as we ha ve seen, is an indissociable mix -
ture of dance, rh ythmic recit ation, and song, again the adepts are the only
ones who t ake part in it , wi th out the support of a single musician, eit her
with his vo ice o r an in strument. U nli ke che pract itioners of the Me vlevi sa-
ma ' , who go int o tran ce as a re sult of their own danc ing bur who need mu-
sic ians to pro vide the mu sic for the dance, the practitioners of the Sadu -
li yya dhikr a re the mu sicants of th eir dance, and thereby are the sole
artisans of their entry int o tr ance.
The Saduli yya dhikr, then, is a perfect exa mpl e of conduc ted tr ance .
Th ere certainly are o chers, and it would see m char the Algerian Rahman-
iyya is one of chem. 76 Bur in certain brotherhoods it is usu al to engage oth-
er musicians. This is true of the H alveri/' who secure a flutist and a drum-
mer for the introductory section of dhikr in particular. It is also true of the
Rifa ' iyya -r hose of Aleppo at least- 8 - among whom the hymn singer
{munshid ) is a lay person, not a member of the brorherh ood . ' 9 Drums -
this rime plaved by members of rhe brotherhood - may also be beaten to
strengthen the rhythm of recitation s or songs. T he drummers then statio n
themselves a little oucside th e circle of those w ho are cak ing part in the ex -
ercise of the dhikr, and the y do nor gen er ally go into tr ance . In th is case,
should we say char sin ce the pract itione rs of the dhikr are not che onl y mu -
sicants of their dance (sinc e they are pa rti ally "mu sicared" ), they can no
longe r be seen as the sole arti sans of their entr y int o tr ance? And would it
then follow char we ca nn ot cons ider the dhikr as bel o ngin g to the ca tego ry
288 CHAPTER SEVEN

of conduc ted t rance' Would the ru le cease to be genera l ? I thi nk nor. In all
t hese cases, the essent ials re ma in: in or de r to ent er into tr ance the subject .
must recite the dhikr, sing, and dance in rhe man ner desc rib ed. T his is
what remains invariable, wha teve r t he bro t he r hood conce rned. It is th is
that defines the dhikr . T he rest - singers engaged from outside, the use of
the drum - should be seen as accessory and consequently as n or inval id at-
ing the rule .
Last, the re is the case of trance linked wi th the man ifest atio ns of fa kir -
ism pr act iced by certain brothe r hoods afte r the exe rcise of the dhikr and
wh ich, through a m isuse of language, is somet imes incl uded und er the
same term. I have al ready stated how we should vie\V t his tr ance: it is dis -
tinct from communion trance, even though it ma y follow it closely, some -
times wirhour a break, and in prac ti ce mingling w ith ir to some ex tent . We
have seen that in eve ry case this trance is associated with m us ic t hat is not
provided by the subject himself (the subject is nor the mus icant of his own
trance ) and that the dance is the expression of tr ance bur not its cause. Un -
like rhe communion trance produced by the dhikr, which is a conducted
trance, rhe "fakirisr" tr ance is an induced trance, as, by rhe same token, are
communion trance (produced by rhe classical sama ·) and profa ne
trance .
This being so, rhe system of the relation of trance to music and dance
a mong rhe Arabs may be expressed schematically by figure 7 .1 : "induced
tr ance" signifies rhar the subject is '·musicared," in other words entranced
bv rhe music provided by those other than himself, and tha t his dance is
moreover the effect, not rhe cause of his trance; '·conduc ted trance" means
rhar rhe subject engenders his own trance, either through dance or
through rhe effect of both his dance and his own action as a m usi canr.

Communion Trance: Between Possession and Shamanism

The induced trance / conducted trance opposition, as it has jusr been iso -
lated, rakes us back to a similar d istinction made in the first p art of rhis
book 80 with respect to possession and shamanism. Viewed in their respec -
ti\·e relations to music , ir will be remembered that rhe possessed perso n
was defined as ·'musicated'' and rhe shaman as a ·'musicant." Communion
trance , shamanic trance, and possession tr ance can thus all th ree be cha r-
acterized by reference to this one opposition, "induced / conducted," bur
\\'ith one provi o: \\·hereas this opposition allows us to dis tinguish be -
tween sha m an ism and possession, in the case of communion tr ance, this
opposition ope r ates from \\·irhin, since it distinguishes between rhe rwo
aspects rhat this caregon· of trance may ha\·e . depending on whether it oc -
curs \\·irhin rhe context of the dhikr or. on rhe contrary, \\·irhin rhar of rhe
classical sama ' . Communion t rance thus seems to lie between po sess ion
Music and T rance among the Arabs 289

Co n d iti ons in which

/
/ '-
-----
\
/ / \
I \
/ .\! ed~"i dhikr I
I ,,,ma' I
I /
\ I
\ conducted /
\
' ______
.........
trance
.,,..,
?/
/

Figure - .1

(ro w hich it is a kin in both the classical samci ' and in its nonr iru alized
fo rm ) a nd sha man ism (to \\·hich it is akin both in the dhikr a nd in t he
}-1e\·levi samci' form ).
A s for profane or emotional trance - tarab - which is ah,·ays induced,
it mu st cle a rl y be pl a ced on the side of possess ion . Mediumship, fo r it s
part . must be set as I have sa id, outside of th is broad opposit io n altog eth -
er.
A.ge neral diagr a m of the rel a tion of tr ance to d a nce and music ma y now
therefore be const ru cted (fig. 7.2 ) by co rr elat ing t he co nt ent s of t he va r-
io u s tr a nce s examined up to 110\1·\1·ith their mode of engende r in g, i.e . ,
either induced or conducted (these t\1·0 1vords b eing taken in the sen ses as-
igned to them earlier with referenc e to fig . 7 .1 ).
Viewed from the perspective of its rela tion to d a nce and music, co m -
munial trance thus seems to oscillate between two poles, th at of posses -
sio n a nd th a t of sham anism . or, rather , it is sometimes similar to the one
290 CHAPTER SEVEN

induced conducted
C

profane trance +
po sse ssion trance +
nonriruali zed +
sa111a'
+
(classical )

religious sa1n{f
communial ritualized
(Mevle,·i )
+
trance

dhikr +
sh amanic +

Figure7.2

and somet im es to th e other. What is th e significance of thi s d ou ble aspect?

Or, more precisel y, if trance 's rel a tion to music and dance is sign ific an t of
its content, if what it is in shamanism is nece ssa ril y different from wha t it
is in possession, how do we exp lain that it can be either on e o r the other in
the case of communion trance? In the first pan of this book, we found th at
possession and shama ni sm are differentiated first by a doubly m arked in -
ver sion of th e relation between the rrance subject and the world of the spir-
its or di vinities; second, by the fact th a t possession trance is in vo lunt ary
while shamanic tr ance is volu ntar y. Sha m anism, as we saw, is a jou rne y
m ade b y m an into the in vis ible wo rl d of spi ri ts; possession is a visit by th e
spi rits or gods to the visible world of men. The relation is oppos ite in di-
rection , but it operates on a footing of equality, since in both cases a reli-
gion of immanence is concerned. Men, spirits, and gods a re all of the same
essence, of which chey represent different a,·atars. some visible and some
invisible. This is why the relation of man ':O the spirits can be eithe r one of
submiss ion or of domination .
In a religion of tr anscendence, like Islam, \\·h ich excludes the possibilit y
of identifying with God, and even more so that of dom inating him, it can -
not be the direction of the relationship chat leads, bv revers ing itself, to
cwo ways of engendering trance, one induced, che other conducted. How ,
then , are we to explain rhat, even here, rwo processes a re at wo rk in op po-
site directions) le appea rs that it is of no great importance whether the re-
lation between the two ,,·orlds, the visible and the invis ible, is one of im-
manence or one of transcendence . Both cases lead to one and the same
alternative : either the relati o n is expe rienced as so mething undergone , or
it is experienced as something willed. In the first case the trance techniqu e
Music and Trance among the Arabs
291
is passive, in the second_ it is active. Let us say, for brevity's sake, that
among the Arabs the dh1kr represents rhe shamanic p ath to comm union
trance, and that the classical sama c represents the possessiona l p at h to it.

An Aside on Christianism
To complement the outline just given on the relation of trance to music
and dance in Islam, we should now examine, in a simila rly systematic
m anner, what happens in those other rwo reli gions of transcendence, J u-
daism and Chrisrianism. 81 The subject is a vast one, much of it lies outs ide
our present purpose, and it does nor present the clear advantage, as was
the case with Sufism, of having a preexisting wri tten theory of these rela-
tions. Let me simply say, limiting myself to Christianism 82 and stay ing
close to the categories we have just examined, that the trance of the Shak-
ers in the United States, and of the Shlustes in Russi a, as it was practiced in
the nineteenth century, would clearly have to be placed alon gside the
dhikr, on the shamanic side, in the category of conducted trances. Like the
Saduliyya, indeed, both Shakers and Shlustes 83 were at once the musicants
and dancers of their own entry into trance. On the other hand, rhe trances
observed among black (and sometimes white ) Americans 8 ' - Baprisrs,
Pent ecosral isrs, and others - would be placed more on the side of induced
trance, and therefore of possession, as a result of the import ant role pla yed
bv musicians (singers, organists, ere. ), and the preacher in the ceremonies .
Bur in this area things appear to be extremely fluid. This is because the
profoundly syn creric nature of these cults, in which European and African
contributions are inextricabl y mingled, complicates nen further an al-
readv complex situation since the ideas of Incarnati on and the Hol y Ghost
nor only weaken the notion of transcendence in Ch risti anism, they also
open up a great man y different possibilities in the area of mystical rela-
tions with God .
Ar the beginning of the twentieth century, there was rhe Welsh Revival -
ist movement, a Christian movement in which mystic trance pla yed a con -
siderable role . 85 T r anee was triggered both by the singing of the partici -
pants themselves and by the action - often described as decisive - of the
preacher. We are dealing with a kind of trance th at was both conducted -
since the participants were musicant s of their own entry into trance - and
induced, bur this rime by means of words, nor song. Unlike the ceremonies
of the Shakers or the Shlusres, chose of the Welsh Revivalists did nor, ap -
parently, include dance. On the contrary, in the famous epidemic of Saint
Virus' Dance 86 (or Saint John's Dance or Saint Guy's D ance ) char swept Eu -
rope and Germ any especially during the Middle Ages, dance was of course
rhe primary sign of trance. Bur was it the cause of this trance or, on the
292 CHAPTER SEVEN

contrary, its effect? The second hypothes is would appear to be the right
one. These dances did not, in fact, occur w ithout mu sic, and since the m u-
sic was pro vided by musicians, the dancers were consequently musicated,
meaning that their trance was induced. This is clearly evident in a drawing
by Bruegel the Elder, 87 the Epid emic Danc e in M oelenbeek, which depicts
a woman falling into a trance as a result of the music being played for her
by a bagpiper. I know of nothing in Europe that is as close to a black Afri -
can possession scene. Except for the costumes and the particular instru-
ments being used, one would think it depicted a ndop ceremony in Sen -
egal. There, we need not hesitate, the subject is a musicared person, and
we are indisput ably on the side of possession. Bur a question does arise .
Whereas in Bruegel's drawing we are undoubtedly in Christendom, we are
not necessarily wi thin Christianism and transcendence.
I shall go no further. These few examples should have sufficiently shown
that the diagram used ro situate the relation of trance to music and dance
in Islam can also be app lied effectively to Christianism.

M USIC AS EMOTION OR AS EX CITATION

Let us return to the Arabs. As we have seen, trance practices amo ng them
are exceptionally numerous. From this great d iversity, however, two
broad patterns emerge : that of the samZi' and of the dhikr . Th e samii.' con-
sists in listening ro music - which is essentiall y singing -for purposes of
triggering trance and then expressing this state in dance. Also in view of
attain ing the state of communion trance, the dhikr consists in reciting the
name of AHah while singing and dancing. These two operations are quite
different. Nevertheless, the y achieve the same result. There is a paradox
here, and it is this paradox that we must now examine.
How, in these two cases, are we to explain the action of the music, or,
more correctly, of the singing since the music is often reduced to singing
an d sinc e in any case, singing always pla ys the essential role? The problem
does nor pose itself in the same way with respect ro dance: in the samii.' it is
the result of trance; in the dhikr it is its cause. We shall return ro this point
later.
Before anything else, however, \\·e ought ro ask what the Arabs them-
selves thought on this subject; and not just the Sufis, moreover, since pro -
fane and religious tr ance, from the viewpoint of their rela tion to music, ul-
timately have a certain number of points in common. The problem
therefore does not only interest philos opher s of religion like Ghazzali; it
also concerns music theorists in general. And as we know, the Arabs have
had some very illustriou s o nes. What were their ide as on the relation be -
t\\·een music and trance?
Music and T ranee among the Arabs 293
Arab Musicologists and the Effects of Music
The first of the Arab philosophers 88 to have dealt with music, al -Kind i19
(who died in A.D. 873 ), thought its effects were of three sorts: those that
"dilated" the movements of the soul, those that "contracted" them, and
those that "calmed" them. Ir was in exactly these terms, borrowed from
rhe Neoplatonisrs, that seven centuries later, Baff, as we saw, presented
the program of his Academie de Poesie er de Musique to Charles IX. In ad -
dition to theories on rhe erhos 9 0 of modes, the harmon y of rhe spheres, and
the efficacir y of numbers , al-Kind1 also proposed a theory on the relation
between sounds and the four elements, the four seasons, and the four hu -
mors. The power of music could thus be explained, according to him,
through a vast set of ps ychological, physiological, and cosmolog ical
causes. 91
AI-FJ°rabf (born in A .D. 872 ), certainly the greatest of the Arab musi -
co logists, also thought that music wa s capable of arousing every passion,
bur for ver y different reasons. Deeply influenced by Aristotle, he refuted
"the Pythagoreans' opinion that the revolutions of the planers an d stars
give rise to sounds that combine harmoniously" (Erlanger 1930; 28 ) and
in hi s huge book on music he did nor make the slightest mention of the
correspondences between sounds and the elements or humors. His theo r y
o f the effects of music is purel y ps ychological and rationalist: "Man and
anim al, guided by their instinct, emir sounds that var y with their emo -
tions, as when the y are jo yful or fearful. Man's voice is guided by other
m oo ds. Ir can express sadness, tenderness, anger. Inversel y, these sounds,
these notes , will give rise to the same passions, the same moods in the
audit o r, and w ill h av e the po wer to ex alt, erase , or calm them " (ibid., 14 ).
Significantly, his chapter on the "effects of melod y" and "their relation to
p ass io ns" form s p art of the discourse he de vot es to "vo cal melodies ." This
amaz ingl y m od ern disc o urse intended "to sh ow how to ad apt the pho-
nemes of a logos to a given melod y and , inversel y, ho w to adapt the notes
o f a melod y to the phonemes of a given logos " (Erlanger 1935; 55 ). Noth -
ing is mo re imp o rt ant to Al-Fa rabf th an a perfect rel ati onship between
me lody and wo rd . "One seek s the mu sical effect, " he w rite s , "in order to
bette r rea ch th e goa l o f p oetr y" (ibid. , 65 ). We find here, rh o ugh vi a a dif -
ferent path from rhar ofNeopl atonism , a fund ament al as pect of what sev-
en centuries later would be the musical theor y of the Pleiade: the union of
mu sic and p oetr y.
The anecdotes about the effects of music that we found circulating in
Paris o r Mil an in the sixteenth centur y were alread y circulating six centur -
ies ea rlier in Baghdad . In one va riant al-Farab1 hims elf is the hero of the
stor y. 9 ' Ir so closel y resembles the stor y of the Milanese lure pla yer re-
co unted by the count of Vintimiglia (w as it perhap s the mode]) ) that it is
294 CHAPTER SEVEN

worth quoting here: One d ay the famous philoso phe r , who was no t o n ly a
music o logist but also an excelle nt music ian, p resente d him self in di sgu ise
at the home of the famous vizie r Ib n cAbbad. T here he sat h im se lf upo n
the ground
amidst the shoes of the people present. It was a gather ing of friends , and
they took him for a jester. He grabbed the instrument away from one of
the singers, an d, having tuned it while he was being mocked, began to
play tunes th at made those listening laugh so heartily that their souls
almost flew away with joy. He then played melodies so melancholic that
the y shed tears and their hearts were about to break with sad ness; then
he finished with a tune th at put them all to sleep on the spot, and left
them as rho ugh dead. On the neck of the instrument he wro te these
word s: "Abu-n- asr of Farah visited your house and then left." When
they awoke, worried and stupefied by the amaz ing things they had
witnessed, the y read what he had written; then they sent people to seek
him out everywhere, bur in vain , which only furthered their
stupefaction.

A litera r r stereo type, of co ur se , that shou ld on ly be tak en for w h at it is,

but which is n one theles s significant: in the Baghdad of th a t time , as in M i-
lan later, what m atte red was th a t music sho uld be expressive, th at it
should mo ve the soul.
But what connect ion does all this h ave with trance, o ne ma y ask? Very
little , even nothing at a ll up ro n ow. We might perhaps say th a t in the sto ry
of ,th e musici an p lung in g his listeners int o th e deepest of sleeps, we brush
up again r trance, in it s hypn o ti c form, just as we bru shed up aga inst it ,
we will recall, under the form of "div in e furor" in the srory of the Milanese
lutenist. E ven so, we h ave to admit that in both cases we do so fr om afar .
We are eve n farther away from ir in the wr itin gs of al-Kindi or al -Farabi.
Tru e, both speak of the effects of music on the "passions" of the so ul , bur
this certainlr does not m ean the y are ref erri ng to trance, by wh ich I mean
that total uphea va l of one's inner equ ilibrium that occ ur s in wajd or tarab .
We know chat al -Fa rabi , who was ver y favorabl y disposed roward Sufis m ,
wrote a book on the samii<, bur it h as u n fo rtun atel y been lost. 93 This
means cha r we shall never know what he th o ught abou t the rel a tions be -
tween mus ic and tr an ce, which is a great pit y.
As for Avicenna, to move o n to the last of thi s great trinit y of Ar ab m u-
sicolog ist s worki ng in the ninth and tenth centuries, he breaks w ith his
predecessors b y n ot m ak ing the slightest mention in his work of any the -
ory of the effects of music in rel a tion to th e cosmos or to numbe r s. 94 Ocher
autho rs after him , as w e kno w, returned to these ide as, notabl y Safi al-Di n
(thirte ent h ce ntur y ), to name only the mo st famous . But w ith the excep -
tion of the Brothers of Purit y, to w hom we shall return in a moment, none
of them was to deal w ith tr a nce.
Music and T ranee among the Arabs 295
Yee one mi ght have expec ted cha t at leas e o ne of chem , al-Hasan al-Ka-
cib (a writer of second ra n k, co be honest, w ho w rote in the earl y elevent h
century ), would h ave tr eate d the sub ject , sinc e hi s wo r k 9 s begins with a
chap ter on tarab. Bue he does not hin g of t he kind. Th e tarab w ith w hich
he is concerned is o nly "mus ica l emotio n" in its m os t int elle ctu alized and
anod yne form , and has very lit tle to do with the dr ama ti c and sometimes
lethal crisis w ith w h ich we are concerned. He also couches on trance,
howe ver, like the write rs mentioned earl ier, with rv, o scories 96 chat w e also
encounter severa l centuries lacer in the writings of Pontu s de T yar d or of
Ronsard . One of them clo sely resembles the T imoth y and Alexan der
a necdote w e examined in the chapte r on Renaissa nce theor ies; the ocher is
a verba ti m ve rsi o n of the sto ry of Pyth agoras and the yo ung T ao rm in ian ,
quoted in the sa m e chapte r. Both were bo rr owed from icom achus. T he
striking thing about al-Katib's use of thes e two stories co illustr ate t he o m-
nip o tenc e of music, howe ver , is that their original pur pose, as we saw,
was to dem o nstrate the opposed effects of the Phr ygian and Dor ian
mo de s. And these two modes are quite foreign to Arab music. Ne it he r
Pont us de T ya rd nor Ron sard was ver y co nvinc ing wi th his rehashed ver -
sion of these old eopythago rean tales. We ha ve already expl ained why.
We see chat for similar reas ons al-Karib is no m o re convincing th an the y.
T o m y knowledge, the only piec e of Arab w ritin g dealing not with Su-
fism but w ith general mu sical the o ry that rakes int o accou nt the relations
between music and trance is the famo us letter of the Brother s of Purit y
(Jk hwan a/-sa/Zi), "a group of philo sophe rs, scienti sts, mathem atici ans,
and litterateurs who flourished at Al-Basra during the seco nd h alf of the
tenth cent ur y" (Farmer 1929; 214 ). T his Epistl e on .\1usi c9 ' (one of a long
series of ocher ep istles ) re, ·ives, developing and amplif ying chem in the
process, all the ide as then cur rent on the influence of music, its effects on
illnesses, on th e harmon y of the spheres, and on th e rel ation of the four
str ings of the lure ("ud ) ro the seasons, elements , and humors. It is a verita -
ble treatise on musical as trol og y. The w riter s of chis epistle 98 defined their
own work as follows : "In this epistle we discuss mus ic; we show also how
notes a nd mel odie s, b y means of h armonious and rh ythmic ass oci ation,
influence th e listener 's soul in the same wa y as me d icine s, pot ions , and
theriacs ac t upon the bodies of animals . We also expla in how the rotation -
al movement of the celestial spher es and th eir fricti on agai nst one anothe r
engenders mel odies as pl eas ing as those of lute string s a nd flute s. The pur -
pose of these melodies is to aro use in the pure and ra tion al, human and
royal souls an ardent des ire to rise up co th ese sires w here the y will be sepa -
rated fro m their bodies, through char separarion known as death ." Clear -
ly, then, tr ance is not the principal preoccupation of this ep istle. It does
nonetheless ha ve a place in it , and a significant one at th at . The last part of
the epistle is an isolated chapter th at has no co nn ect ion w ith anyt hing p re-
296 CHAPTER SEVEN

ceding it and that is entitled "The Di verse Effects of Music. " 99 It deals ex -
clusively w ith the Sufis and wajd . Sever al anecdotes are quo ted, all very
similar to tho se told by Ghazzali, in which Sufis are seize d by trance and
commit their souls, or faint upon hearing one or anothe r passage of the
Koran or of sung poetry.-Although the title of the chapter attributes these
effects to melodic modes, the se modes are never specified. The sung or re-
cited words, on the other hand, are quoted meticulously. The so ul rejoices
and is exalted, we read, "when it listens to descriptions that correspond
\\·ith the object of its desires and to melodies th at are in harmony with the
object of its delectation." Given the necessary union of melody and words,
it is the meaning of the words, above all else, that is really important. This
being so, no explan ation is provided for trance as a particular effect of
music . It is presented as a fact, and th at is all . "May God aid yo u, oh
brother, to understand . . . his hidden secrets." This is how the letter
ends .
Our excursion among the Arab theoreticians, howe ver interesting it
ma y have been, has not reall y solved our problem. We will thus return to
I t.

SamZi ' and Emotional Tr ance

As \\·e have seen, two broad models emerge from the diversity of trance
forms among the Arabs, the samZi' and the dhikr : two practices that could
not be more different from one another (what could there be in common
between listening to singing, seated and impassi ve, and sing ing oneself,
even howling, wh ile flinging oneself about like a madman? ) but th at both
paradoxically result in the same thin g, namel y "commun ial" trance. Let us
examine them again more closely, one after the other.
We shall begin with samZi ' . "Audition" (samZi') of a song or of a chanted
reci tation with religious content triggers trance (wajd ), which can be
either expected - in which case it is samZi ' ritual, which results in dance -
or, on the contrary, unexpected - in which case it is crisis or fit, which
leads to dea th or madness. By means of what power does this singing trig -
ger trance or crisis? Ghazzali tells us clearly, and all the anecdo tes we ha ve
read confirm his statements : it is through it beauty - the quality of the
vo ice in particular-on the one hand, and on the other through its mean-
ing, which has no moving quali tie s in and of itself but owes them to the
subjecti\·ity of the auditor. It moves him because it is in perfect accord with
his inner state at that moment , which is that of desiring and seeking God .
He experiences the situation as an abrupt meeting - the very meaning of
the word wajd - a revelation, an illumination. Hence the emotional aspect
of the process, an aspect all the more marked because this meeting takes
place in a context itself imbued w ith passion , since love of God, which in-
Music and T ranee among the Arabs 297
spires this desire and this search, is always more or less im p regna ted w ith
erotic ism - whether it is sublimated or not changes noth ing.
In profane life tarab corresponds to wajd -=religious trance associat ed
with the "spiritual concert" that sama <represents. The story quoted earl ier
about the poet seized by tarab while listen ing to t he singing of th e celeb rat-
ed Jamila shows that in this case it was also the beauty of the voice and rhe
erotic context - profane this rime - that were at work. Whether tarab o r
wajd, the trance manifests itself in the same ways : by dancing; by that odd
form of behavior ( tamziq ) that consists in rearing one's garments; by a
dead faint, or possibl y by all three in succession. Like wajd, tarab can be
fat al if it is triggered by an unexpected encounter with a sung text: witness
the story of the wife taken away from her homeland by her husband and
who, upon hearing some lines relating to her situation, was seized by tarab
and died. Here again it is the meaning of the words and their d ramatic
content, related to the listener's subjectivity, rhar are responsible for the
trance.
Let me sum up. Both the spiritual and profane concert can give rise to a
trance that will in both cases manifest itself in the same way. Unexpected
audit ion of a religious song or unexpected audition of a profane song -
both can trigger a fatal trance. \Vajd and tarab, religious trance and pro -
fane trance, thus seem to be totally homologous, except for their content.
\'tie can thus say, that from our perspective, we are dealing with nothing
more than two forms, one religious, one not, of one and the same tr ance.
We can say that tarab is wajd minu s references ro God; or, reciprocally,
that wajd is tarab plus reference s ro God. Seen no\,. \\·ith respect ro the ir
relation ro music and dance, wajd and tarab are al o homologous. Both
are "induced" trances in which the subject is "musicated," song is the o per -
a rive factor, and dance is the expression (nor th e cause ) of trance. In both
cases it is esthetic and passionate emotion that triggers trance. In short, in
both cases we are dealing w ith "e motional tr ance,'· and it is this emotional
trance that we will examine in the remainder of this chapte r.
One observation may perhaps be useful here. Belie\·ers w ill doubtlessly
consider th at ro merge wajd and tarab under one he ading in this \\'a)' is
equivalent ro remo\·ing God from religious trance , and that to do so is ab -
errant . since it is preciselv an encounter w ith God that produces it. Th is is
nor the case. Whe ther one is a bel ie\·er o r not, the fact is that wajd and
tarab are triggered and manifest themselves in the same \,·ay, and their re-
lation to music and dance is the same . The revelation can thus be that of
God, of beaut y, or of truth; what counts, ultimately, is the emotion it
arouses and the trance that ensues . Thi s being sa id, belie\·ers are free tO
think that when it is religious, emotional trance is radicall y different from
w hen it is profane, and others ma y think rhe contrary. The two positions
a re equally compatible with my interpretation of the facts .
298 CHAPTER SEVEN

This practice of emotional trance among Arabs brings us to a situ at ion

we have nor yet encountered in this book. We must now therefore tall.<
about music al emotion and examine its particular faculty of causing the
Arabs co become entranced.
In his preface to Erlanger's great work (1930; ix ), Carra de Vaux, afte r
first having stressed the essentially melodic nature of traditional Arab mu-
sic, observes: "With such restricted means, music produced an enormous
impression upon Middle-Easterners. A simple disrich accompanied by the
lure, a few introductory notes suhg by a beautiful voice, possibl y a vo ice
with the slightl y raucous and guttural timbre as they love it in the Middle
East, was enough to throw the listener inco a state simila r to that of ecsta-
sy ; he quivered, wept, fainted, he thought he was go ing to die. Arab litera-
ture is full of anecdotes bearing witness co chis hyperesthesia of the musi -
cal sense ." Ir is indeed hyperesthesia that is in volved, it seems, at lease
relative to wha,c we late-cwentiech-century Europeans see as the norm. Bur
to what should we attribute ic) I cruse everyone will agree we cannot see it
as some natural and innate quality of che Arab temper ament. We are
clearly dealing with a sociocultural phenomenon (even though one could
just as easily interpret our rel ati ve indifference to music as resulting from a
veritable form of anesthesia of our musical sensibilit y ; in which case chis
would then be our culcu r ally specific characteristic ). This hyperesthesia,
we have said ever since the beginning of chis cha peer, is largel y regressing
in our day in the cities. This is proof, if proof were needed, chat it does in
fact seem from a certain kind of upbringing, tied to a certain type of soci -
ety.
Moreover, although this musical hyperesches ia is certainly much more
widespread among the Arabs than elsewhere, there are other examples of
it in the \\·orld . In our day, in the realm of Western pop music, che behavior
of fans electrified by their idols (such as Janis Joplin or El vis Presle y), be-
side themselves, yelling, and even sm ashing up che audito rium, may be
seen as a:cultural (or countercultural ) equivalent of the re ars, fainting, and
garment rending of the Arabs.
An even better comparison is co be found in what Rousseau 10 0 had co say
ab out che effect of the "R anz des vacht s" upon Swiss soldiers. When he
talks of this "air so dear co che hearts of the Swiss chat playing ic in their
Troops was forbidden on pain of deach, because it made those who heard
it weep, desert, or die , so ardent a desire did ic arouse in chem to see their
country again," is he nor in fact describing a completel y typical case of
tarab? For, one will recall , che wo rd tarib, meaning subject to tarab, is
appl ied to the person "moved by the memor y of his homeland" and "lo ng-
ing" to see it again. I have just referred a second rime to the story of the
woman who is taken far away from her kin by her husband, and who dies
suddenly upon hearing someone sing verses about homesickness. We may
Music and Trance among the Arabs 299
be sure that Isfahani, who recounts this srory, would ha ve found nothing
at all astonish ing in a Swiss soldier who passes from life to death simply
upon hearing the "Ranz des vaches." He would have perceived no "hyper -
esthesia of the musical sense" in the event. For him, the fact that an emo-
tion of a certain type could be so strong as to bring abo ut instant death
was part of the natural order of things, just as much, for instance, as see-
ing a caliph having his best friend's head cut off in a moment of violent an -
ger. It is characteristic of all cultures that they should repress or, on the
contrary, encourage, the expression of emotions, and that the y sho uld de -
cide which among them should or should not be valued. The person who
could explain why Arabs have so strongly developed their musical emotiv -
icy would cast light on wha t is certainly a fundament al aspect of their civil-
ization. The fact that for them song (ghina c) and trance (tarab ) are so
closely associated c;,at the two words can be used synonymously is cer -
tainly indicative of a very deeply embedded feature of this civilization. In
che absence of further information, lee us simply rec all that Arab singing is
aimed particularly, and above all , at moving, or rather at overwhelming
the listener, and that the latter is in turn encouraged by his upbringing to
cult ivate this emotion, sharpen it, exacerbate ic as it were, and not only
give it free expression but make it explode, and express this explosion in
the most spectacular way. Such conditioning evidently predisposes some -
one to trance. It does not, however, explain ir. In order to do this, I w ill
venture a hypothesis. ,
It is clear that for Arabs music possesses the emotional powe r w ith
which we are concerned here only insofar as it is associi)ted with words,
which is to say wi th mean ing . At the sa me time, however, it is because
they are cast into a musical form that the words have this impact, whic h
they would lack on their own. Two complementary facto rs are therefore
ac work : on che one hand meaning, on the ocher the sensitive form of the
meaning. One is nothing without che ocher. Only melodies that have a
meaning, melodies to which "something exterior to them corresponds,"
al-Farabf says, are perfect (Erlanger 1935; 95 ). Thi s was also Plato's opin -
ion, for whom purely instrumental music was merely an aberra tion, as we
know. But although the meaning can sometimes be cultu r ally specific
(when it relates to the inaccessibility of God, for example ), someti me s uni -
versal and thus natural (when it expresses homesickness ), its sensitive
form (in other words the musical form it is given ) is always cultural. It is
therefore only insofar as it refers ro culture that the musical form affects,
and even overwhelms, the listener. Such a cultural reference thus suddenly
confronts the individual w ith what formed him, what fashioned his sens i-
bility, with everything that is consequently anterior ro him, all he has al-
ways known, and all that transcends him. This experience confronts his
fleeting, imperfect, unfinished individuality with permanence, wi th com -
300 CHAPTER SEVEN

pleteness and with ontological plenitude. It makes him aware in the most
sensitive way - because it is done through his senses - of the existence of
two opposing realities. his own and another, o ne that is both and at the
same time close yet in opposition to his. The dichotomy between wh at he
is and what he is not - to which he obscurely asp ire s-i s then felt w ith
heart-rending ac uteness. T ranee, or rather crisis, is undoubtedly due to
this acute and intense impression of being inwardl y torn apart; it consti-
tutes a response to an inner state that has become untenable, and for
wh ich it is an outlet. Two paths are thus possible. Either this contradic-
tion between the two present realities cannot be resolved and the crisis,
wh ich cannot blossom into trance, leads to m adness and death, or the oth -
er reality does not merely erupt into the re alm of consciousness bur totally
invades or occupies it. In this case it is illumination, which can be so br ief
as to merge with the crisis - collapse, swoon-or, on the contrary, can be
of a certain duration and manifest itself either by the rending of gar -
ments 101- a highly stereotyped behavior pattern - or by dance.
Possession tr ance, for its part, is associated with a division of conscious -
ness in rel ation to a general conception of the structuring of the person in
which the latter is thought to possess several "souls." Emotional trance is
not linked with this ideology, but in its religious form it is clearl y related to
the idea of God, which is also that of transcendence. This tr anscendence
could well be but one specific aspect of a more general rep resent at ion of
the world, which conceives the latter as irreducibl y dualistic. Pluralit y of
soul or duality of the world, the trance that results from them is in both
cases the sign of a certain experience of being torn, and chis possibly con-
stitutes the common denominator of thes e two forms of trance chat in och-
er respects are so very different.
This being so, and regardless of whether my hypothesis is acceptable or
not, one thin g must nonetheless be remembered: for an Arab music has
the power of inducin g trance only because it is a vehicle for wo rds, and be-
cause these words are charged \\·ich meaning. As Rousseau would say, 102
"he must understand the language spoken to him before what he is being
told secs him in motion," which proves chat it is not "the physical power of
sounds" chat is at work here. Thu s we find chat in chis respect what we
said about the relations between music and possession in the first part of
this book also applies to emotional trance .

Dhikr and Excica ti onal Trance

Whereas the ritual of sama ' can be said to pave the way for trance through
purely psychological means, the ritual of dhikr app ears to do so mainly
ch rough physiological means.True, the repetition of the name of Allah, or
substitutes for it (dhikr ), is an exercise chat occupies the mind , but in the
Music and Trance among the Arabs 301
collective dhikr intoned aloud, which is th e one t ha t interes t s us, it is the
in vol vement of the entire body in very viole nt p rac ti ces th at m anife stly
constitutes the most important part of the exercise . These prac tice s esse n-
tiall y ha ve two aspects: half -spoken, half -sung recitation, and da nce, t he
rwo being coordinated by rhythm, which is very heavil y stressed. All this
has already been described, so we need not go ove r it again. Let me just re-
call rhe central characteristics.
Contrar y ro what usually happens in both speech and song, in which
vocal emission occurs onl y while breathing out, here the vocal cords vi-
br ate as much during ex halation, with only the briefest interrupt ion at the
moment when the mo vement is reversed. T he voice is guttural. The rau-
co us and racking pa nting resulting from it recalls the sound of the back-
and-forth movement of a saw. In fact, this chant ing appears co be some
sort o f vo cal exercise in which the consta ntly forced modulation of the
bre ath pl ay s a ver y important role. It seems likel y that breathing in such a
w ay- sometimes for as much as several hours on end - is likely ro pro -
duce cert a in ph ysiological effects. To my knowledge, they have not been
studied scientificall y. It seems justified , however, ro assume that the ensu -
ing pulm o nar y hyperventilation is likely co produce disturbances chat
might be of the order of a "dizz y spell ," let us sa y. The hyperactivit y of che
voc al cords, moreover, in conjunction with overscimulacion of the hearing
system, which necessaril y results from vibrations in the throat that are in -
tern ally transmitted, also undoubtedl y contribute ro modifying the vascu -
lar and neurological balance of the encephalon. As for the dancing, which
co nsi sts o f a sequence of fairl y elementar y figures, it coo is ch aracterized
by an altern ating m ov ement consi sting of swa ying the entire bod y from
right ro left a nd left ro right while the head oscill ates fro m one side ro the
o cher. H ere aga in , it seems likel y chat such m ov ement s ca n in the long run
ca use di sturb ances of equilibrium and res u lt in the "spa ti al dis o rientation "
co which Zemp leni refers with respect co the nd bp .
T he dhikr cer em o ny can thus be seen as a corpore al tec hnique aimed at
bringing a bo ut ch ange s in the indi vidu al's ph ysiolo gica l equilibrium char
w ill lead to cert a in perturb ation s. Th e con vulsi ve aspect , so o ften de -
scribe d as being typic al o f chis kind of trance, natur ally m ake s one think of
so methin g like o rga sm . 10 3 Experienced in a state of consciousne ss ent irely
occ up ied with the ide a o f God , \\·ho se name is repe ated throughout, any
resultin g di zziness, a ny convulsi o n , naturall y co uld be interpreted , given
rhe mystic ex alt ation in volved, as an abrupt encounter wirh God . This be -
ing so , and whether the ps ychoph ysiological proces s char occurs is evident
o r no r, chis tr ance is nonetheless attained largel y by me ans of the hyperex -
cirario n o f a sec of neuro vegetarive functions. I therefore propose that we
cal l ir "excirational" tr ance, ro distinguish it clearl y from rhe one attained
throu gh the classic a l sama ' , which I h ave called "em otion al" trance.
302 CHAPTER SEVEN

The collecti ve dhikr thus ap pe ars co be a very specific singing and danc-
ing techn ique-in ocher words a musical technique in the global sense of
the term "mus1c" -pr acnced with the aim of achieving a certam effect
up o n oneself, namel y chat of triggering one's own trance. To call this
trance attained by the pr actice of dhikr "excitat ional ," as I ha ve just done,
am ount s co advancing a theor y, howe ver elementar y it may be, on the "ef-
fects" of che music concerned. Before going any further it would be wo rth
ask ing at chis p oint if the Arabs themselves have ever adva nced an y theory
on chis subject.
The musicologists of 'che golden years-al-Kindi , al-Farabi , and Avi-
cenn a- who as we have seen, are all silent on the effects of music in the sa-
ma c, are all equall y mute on the subject of such effects in the dhikr, for the
excellent reason chat the collecti ve dhik r, w hich does not date back be-
yo nd the twelfth centur y, was unknown to chem . The same applies co
Ghazzali and all those w ho wrote on the sama ' before him .
Ther e could, however, be a number of Sufi rexes , which, when set
alongside the oral evidence given by dhikr pr actitioners , m ay enable us to
formulate a theor y of th e dhikr . Thi s the ory, accordin g co B. M auguin , ' 0 '
could be for mu la ted as follows : "In a z ikr, the function of the music, both
vocal and instru ment al, is to act iva te the va rious centers in the human
body. It is no t perfection of fo rm char is being sought; the only thing that
matters is the circulation of acous tic energy, the perception of a b as ic
rhythm, an d awareness of vibrator y phenomen a which, when intensified ,
are a lone able to eliminate the superfici al level of the self. Ir is as though
the ancient masters would have known - a knowledge w hose origin is dif-
ficult to explain - th at cert ain ner ve endings import ant co the psyche are
located in th e thro at. In this way, breath ing, acce ler ated and subjected co a
musical rh ythm, becomes a powerful faccor in mind co ntrol. " Th e impor-
ta nce given co the "ner ve endings in the thro at" rends to indicate char
chanc ing- w ith its rather pa rticul ar technique - would rend co be the
pr ime facco r in the process. Bur wha te ver the case may be (for one would
like co kno w more o n these ideas and their origin s), we are presented with
a neurophysiological theory o n the effects of music in the dhikr.
Depont and Coppolani, in their by no mea ns recent description of this
ritual (1897; 156- 57 ), asso ciate rotation of the head (wh ich is swung
from side co side then lowered onto the che st ) w ith the "raucous sounds"
of the vo ice, compressed by "cer ebro -spin a l congestion"; "the blood beats
ha rder and harder in their temples, the participants col lap se he av ily, roll
on the gro und ," the y wr ite , and chis is "ecstasy."
Garder (1955; 20 3 ), co who m we owe the most thorough published
study on the dhikr co d ace, chinks that it should be regarded (I am summa -
rizing a great deal ) as "a sec of verbo-mocor and respiracor y means " act ing
"on che cerebro -spinal ner vou s system " in order co in some way bring the
Music and Tr ance among the Arabs 303
"conscio us ... into ha rm ony ... w ith the subco nscio us," thus impos -
ing the "desi red m onoideism" on the one as on r:1e othe r. In a note refer-
ring to certa in psych iatr ic re search, he emphas izes, as support for his the-
ory, the interdependence of the vegeta tive and the cerebrospinal ner vo us
systems. So although Carder's personal position is far from be ing th at of
an unbeliever, his theor y of the dhikr as trance technique, as we can see,
clearly allots the largest place to physiology. ' 0 '
What are we to make of these va rious theo r ies? Fir st, there is no avoi d-
ing the fact that none of them - and my own is no exception - is scient ifi-
cally well founded, and that they all p resume, with an adm ixture of intu-
ition and common sense, that if the dhikr - a practice expressly intended
to induce trance and in most cases succeeding in do in g so - brings into ac -
tion ~uch a specific corporeal technique, t his is because the physiological
upheavals it produces have the power to pro vok e that trance. H ow so?
That is what we need to know, for this is precisel y the point on w hich we
are ignorant. This does not, howe ver, invalidate the hypo th esis. Let us
therefore accept it and mo ve on to invest igate its implications.
In the dhikr we may have, at last, acquired proof that a certain musical
and dance practice - a very ritualized one to be sure, but this does not
matter - has the power of producing trance unaided. If music is indeed ca-
pable of such effects in this context , rhe first thi ng we must then ask our-
selves is if it might not be capable of doing so in other circumstances, and
if we ought to consequently revise all we have said thus far in this book
about its relation to possession. Let me answer right away that this is nor
the case. According to my hypothesis, the condition sine qua non for mu -
sic to act as it does in the dhikr is that it be a practice on the part of th e
trance candidate, so mething it never is (with one exception that does nor
invalidate the rule, as we saw ' 06 ) in the case of possession. Music in the
dhikr operates as a corporeal technique, nor as an organization of sounds,
hence its characteristic of being strictly indissociable from dance. Th e
dhikr cannot, then, in any case, be seen as an example that would demon -
strate that sounds in and of themsel ves have the power of producing
trance .
Ta king account, then , of the fac t that it is the practice of music and nor
the music itself - not the mess age but the ac t of emitting the message -
that is responsible for trance, we must now go on to investigate w hether
this practice, which certainl y is a necessar y condition for the ritual to in-
duce trance, is also a sufficient condition; in other words, if, in order to go
into trance, it is enough to practice dhikr in the correct m anne r. Or again,
to pose the problem in ano ther way, whe ther the practice of dhikr induces
trance automatically, as has been claimed . For this is how it is generally
conceived, and it is precisely this mechanizat ion (if one may so put it ) of
trance for which the dhikr is critic ized. ' 0 ' In this case we are once again
304 CHAPTER SEVEN

faced with a situation entirely comparable to the one we encountered ear -

lier in the case of possession ·, 10 ' where the "right tune " (the mus ical motto )
is often thought to have the power of automatically triggering trance. In
fact, as we saw, this is not at all the case. And the same is true, as we shall
see, of the dhikr.
Let us begin with the most astonishing case, which is mentioned by Tri-
mingham (1971; 211 ) at the end of his description of the dhikr among the
Mirghaniyya in the Sudan. Th is brotherhood practices a collect ive dhikr
that confo rm s in all points to th e general picture given in thi s book , in-
cluding the vocal emission, since in the final stages, T rimingham tell s us,
the words are reduced to a barking noise from the chest, or to the r asping
sound of a saw. And then he very unexpectedly adds that this brotherhood
is in no way seeking to produce "ecstatic phenomena." 109 Up to now, all
aspects of the execution of dhikr were seen as answering to a very precise
objective, that of inducing trance. Suddenl y we are faced w ith an example
of dhikr that does not ha ve this aim at all. Thi s is in itself surpr ising, bu t it
provides the answer to our question: if dhikr, practiced here and there in
the same manner, sometimes leads to trance and sometimes does not, this
means that it does not induce trance automatically; in othe r words, some-
thing else has to be added co it. But what) Before we tackle this question,
let us take another look at the evidence.
That dhikr should be routinely practiced without any intention of trig-
gering trance certainly is a pa radox - and one that raises an interesting
problem. At the same time , the facts show that the sheer practice of dhikr
is not enough to induce trance automatically, or at any rate that it does so
only given certain conditions . This is corroborated by another piece of
evidence that is probably an everyday occurrence but that I m yself wit -
nessed only once, 110 namely that those t ak ing part in "classical" dhikr -
th at is, one intended to induce trance - do not all necessarily become en -
tr anced . Some do, others ma y not. Yet all, apparently, are performing the
same exercises in the same wa\·. Moreo\'er, it is well known that the effects
of dhikr do nor manifest themselves in rhe same way in every brother-
hood. Among some, the trance to wh ich it normally leads is unb ridled,
convulsive, and paroxysmic; among others, it is relati vely calm and con-
sists in little more than a gentle and ephemeral daze . This is the case, nota -
bly, among the Derqawa d'al -Jadida in Morocco, 111 whose exercises -
swaying of the bod y, roaring, etc. - are just as violent as those of the
<lsawiyya, for example, yet never lead to comparable violence in the
trance state . Fin ally, one other aspect of the m atter: we saw that in posses-
sion the neophyte is the one who responds almost w ithout fail to the sum-
mons of the "right tune, " whereas the officiant remains impe rvious to it, or
only submits to it with difficult y; the same is true in dhikr, in wh ich the
ordinary adepts go into trance whereas the shaikh does so only rarely . 112
Music and T ranee among the Arabs 305
T hese few observations shoul d have suffic ient ly demo nstr ated that the
dhikr, wh ich usually lea ds to tr ance, ·ca n also not lea d to it , and that ,
moreover, the ma ni fest ati ons o f the tr anc e are va r iable. As a hyp othe sis, I
suggested ea rlier th at the musical prac tice of dhikr is respons ible fo r the
triggering of trance; now I find myself forced to say that the same pr actice
is not always followed by the same effect, wh ich means eit he r th at m y hy-
pothesis is false or tha t it is inadequate. If the sa m e cause does not always
produce the same effects, it necessa rily fo llows that either the rela ti on is il-
lusory or that it has been distu rbed by some ot her eleme nt , or aga in th at
the cause in question is not the only one at work. It is diff icult to acce pt
that the hypothesis is illusory, because this wo ul d mea n t he ge neral hype r-
st imulation provoked by this "exc itational" practice is unrelated to the
triggering of trance. Why then is it systematica lly sought afte r ? Let us con -
sequently remain faithful to our hypo thesis but try to improve on it.
To begin, however, I would like to rejec t o ne possibl e objec tio n that
might be raised, namely that in the field of human behav ior no law can be
verified other than with a fairly wide margin of uncertainty, and that in
this case, it would be futile to continue ou r quest : since the context inevita-
bly varies from one brothe r hood, session, or man co another, it follows
that the effects of this pract ice would also vary. My answer is that we are
not dealing w ith such a case. If the pract ice of dhikr among the RifiFiyya is
expressly intended to produce trance, and they consider that the cause -
and -effect relation between the two is self-evident, and if among the
Mirghaniyya, on the contrary, the same dhikr is practiced ,vithout any at-
tempt to induce (and, to be sure, attain ) the trance state, we are faced wi th
a problem, whether we like it or nor. This practice, as I said, is above all a
corporeal technique . As such, in and of itself it should have the same phys -
iological impact, broadly speaking, on everyone . The problem, therefore,
is the following: either it produces trance, as in the first example, in which
case why does it not have the same effect in the second) Or else it does no t
produce trance, as in the second example, in which case why do t he par -
ticipants in the first practice it, since trance is thei r sole aim)
This contradiction is resolved if we accep t tha t th is excitat ional prac t ice
is- in the context of the dhikr, naturally - a necessa ry condi tion for tri g-
gering trance but is not sufficient in and of itself. In other wo rds, the re is
another cause involved, not phys iological this t ime - since the physica l
practices of dhikr remain the same in all cases - but psychological.
Let us accept, then - and this is my first hypothesis, which of course de -
mands verification - that the prac t ice of dhikr gives rise to phys iological
disturbances, but let us add that the latter constitute no more than pr ep-
arations for trance. Our second hypothesis will be tha t these prepa r at ion s
only bear fruit if they are conceived as a pr elude to trance o r , an d this
amounts to the same thing, only if trance is seen as the no r mal end result
306 CHAPTER SEVEN

of these d isturbances . T ra nce wo uld then on ly occ ur wh en two cond itions

are met: first, that one pract ices (to be sur e) dhikr; seco nd that one per-
ceives dhikr as something that res ults in tra nce. In oth er wo rd s, if the
model demands that the dhikr engende r trance, then it does (Ri faciyya ); if
the model does nor require trance in its program, then trance is not engen-
dered (Mirghaniyya ).
What a useless truism, one might say. By no means. T his inter pr etat ion
of dhikr as functioning to generate trance in fact proposes that the pract i-
tioner attains the trance scare only because he conforms to a model sug-
gested to him, in ocher words, because he identifies his behav ior wit h th e
model. In order for the disturbances brought on by the phys ical practice of
dhikr to lead ro trance, it is necessary that the intention ro transform them
to trance should exist somewhere in the practitioner's conscious or uncon -
scious mind. Transition into the trance scare would thus be a del iberate act
on his pare - wh ich does nor necessaril y mean char it is a conscious act.
The overall process could thus be described as caking place in two stages,
even though they may not be distinct from one another in reality . The ef-
fervescence caused by the practice of the dhikr would lead the individual
ro such a state of instability or inner confusion chat it would take very little
co topple him into that particular stare called trance . It would lead him co
the threshold of this stare, bur, without any intention on his part, it would
be insufficient in and of itself to effectuate the passage . An intention co en-
ter into trance is indispens abl e to chis con version of the quantitative into
the qualitative, this phenomenon of crystallization that cha racterizes
dhikr . The musical practice of dhikr - song and dance - could thus be seen
as comparable to the operation requ ired for bringing a liquid to its satura-
tion point in order to cause a change in its state; in both cases a further and
quite distinct operation is necessary for the change co cake place . In the
case of trance it is the will - be ir conscious or unconscious - to rake the
Step.
Thus entr y into trance, in dhikr, would be a deliberate act, bur one ne-
cessitating long preparation. T he hypothesis enables us to solve economi -
cally the problem raised by the existence of dhikr practiced without a
quest for trance. The reasoning behind chis is as follows: w hen practiced
without any intention of going into trance (M irghaniyya ), dhikr does no r
in face induce it; wh en it does trigger trance (Rifihyy a ), it is because an in-
tention to go into trance was already present. This hypothesis also re-
solves the problem r aised by the fact chat the practice of dhikr does not
ha ve the same effect - among the same people and dur ing the same cere-
mon y this time - on the shaikh as on the adepts. As a general rule, as l
said, the latter (or the majority of them ) go into trance; the shaikh doe,
not. Th is is because the ad epts had an intention co do so whereas the
shaikh did not. (This does not , of course, preclude the possibil ity char the
Music and Trance among the Arabs 307
adepts may have the intention of going into tr ance and yet fail to do so, or
th at the shaikh ma y go into trance without ha vin::;intended to do so at the
outset. We cannot always do as we wish .) Logic , clearly, is on our side. But
the explanation ma y very well seem somew hat simplistic to the uncondi-
tional supporter s of the spontaneous trance theory. And it does undoub t-
edl y pose the problem - hackne yed but inevitable-of the sincerity of
trance. I have menti oned thi s aspec t before. We must return to it no w.

W illed Tr ance , True Tr ance ?

Ghazzali, in the seco nd part of chapter 2 of his "Book of the Right U sages
of Audition and Trance," poses this problem in very clear terms. For him,
not cnly can trance be sought deliberatel y-i ndeed , this is the very defini-
tion of sama <- bur the actu al entr y into tr ance ma y be feigned. This does
not mean th at the trance is insincere; ever ythin g depends on the purp ose
for which it is feigned. Since tr ance , for Ghazzali, is a form of beha vior
th at is learned - and practiced in accordance with "right usages "- it is
only na tur al that in order to learn it , one begins by simul atin g it. There are
two sons of wajd, he tells us (McDonald 1901, 730 ), o ne th at is spontane-
ous and one that is fo rced , the latter being in turn divide d into two cate-
gories: one that is blameworthy (because it is hypoc ritic al) and one that is
praiseworthy. "And this is why, the Apostle of God comma nded him who
did not weep at the read ing of the Koran that he should fo rce wee ping and
mourning; fo r the beginning of these States is some tim es forced w hile
their ends thereafter are true. " orhing is earned "except by effort an d
practice," he adds, "the reaft er it becomes nature through custom" (ibid . ,
731 ). At the end of his book, in the pa rt that expressl y sets our the "right
usages," he returns to this point and states, first, that one "shou ld not rise
[in order to dance], or r aise his voice in weep ing while he is able to restrain
himself" (McD 1902:8 ) and second, that "if he dance and force wee ping,
that is allowable whenever he does not intend hypocrisy by it; for fo rcin g
weeping induces grief and d anc ing is a cause of joy an d liveliness."
Ghazzali is not speaking of dhikr - or at least not collective dh ikr - but
of sama c. Let us say that, in our terminology, he is talk ing about emotional
and not excitational trance. In emotional trance , the relation between
trance and that wh ich tri gge rs it- the emot ion - is evidently less close
than in excitational trance, in which the exci t ati on th at tri ggers the tr ance
appears to be at one and the same time its cause and its symptom. For
Ghazzali, emotional tr ance was a learned form of beha vio r, which one ac-
quired by first beg innin g to simulate it. He could have had even gre ater
re aso n to think the same in the case of excira ti onal tr ance , in w hich not
only does one mime the excita tion of tr ance before it happens , but one
also chants the name of God in such a way th at it is difficult to know
308 CHAPTER SEVEN

w hether he is being invoked ro manifest himself or whethe r he is being ad-

dressed because he is alread y th ere. In both ca ses , however, one must stil l
understand how, by what process, the transition from simulated to real
tr an ce rakes place.
Life pro vid es man y examples of emot ions that begin as feigned or
imaginary but then become re al, often against one's will. One can frighten
oneself in pl ay, then end up being genuinely terrified. Th e same is true of
anger, grief, and eve n joy, although this is rarer and more difficult. At a
certain point, one no longer kno ws quire w hat the truth is, but by the end
on e's consciousness is undoubtedl y totall y overwhelmed by the emotion.
On the basis of w hat? Of nothing 1 It is more difficult, howe ver, rounder -
st and how th e same mech an ism could work for trance. Perhaps because
one is in _the habit of thinkin g of it in all-or-nothing terms, unlike emo-
tions, whic h we custom aril y co nceive as havi ng va rying degrees of inten si-
ty; perhaps also because we have had no personal experience of trance. I,
anyhow, ha ve not . But if trance is a state of gre at em otion , accompanied
by certain ment al repre sentatio ns and certain physiological disturbances,
then it is conceivable th at disturbances of the sa me orde r, artific ially pro-
voked by exerc isin g dhikr, and accompanie d by the same represent ations,
would gr eatly aid th e transition in to the trance state.
In trance, be it com muni o n or possession tr ance , the indi vidu al's beha v-
ior is so radically different fro m his no rmal be hav io r it seems as though
some othe r person we re involved: he was stand ing- he falls; he felt
pain - he is impervious ro it; he was clothed-he rears off his garments. It
is fairly easy to accept thi s substitu tion if on e th ink s of it as in vo lunt ary: if
th e individual who is its cheater undergoes it , if it rakes place aga in st his
will, because it is stro nger than he, because he is bo rne alo ng by an emo-
tional rorment that has deprived him of all control over himself. Ir is r at h-
er difficult to accept that he may be an accomp lice in th is change or, even
more than he is actually the instigator. One thu s rend s ro read int o his be-
havior an insinc erity, a duplicity that is incompa tib le wi th the re alit y of
the substitut ion . In sho rt , one thinks th at rhe person is an imp os ter. Thi s is
appare ntl y an error. Tr ance can be w illed. E veryo ne will agree that dhikr
involv es a deliber ate attemp t ro attai n the trance state . By the same token,
we should not be aston ished ro learn th at it is also, p arti ally and for good
reasons, a simulat ion of trance. Thi s is an appa rent contradiction th at we
have to rake into account: tr anc e can be volunta ry, bur this does not me an
it is necessa ril y insincere.
Let me add that if collect ive dhikr is a deliber at e attempt to provoke
trance, "so litar y" dhikr appears to be a no less delibe rate attempt to
achieve ecstas y. On thi s point, howe ve r, the Sufi s themselves, alre ady di-
vided as to the rel ative va lue of the two forms of dhikr - collecrive or soli -
ta ry- are even m ore divided on the question of w hether dhikr invo lves a
Music and Trance among the Arabs 309
deliberate quest for or a spontaneous realization of trance. Al-Hallaj, as
we know, insisted that trance must never be actively so ught, but that it
should always be a gift from God - a conviction sha red , late r on, by Saint
Teresa of Avila (though without the same results , fortunately for her ).
Let us return to music. The keystone of collect ive dhikr, if o ne may so
express it, is a particular musical practice. If the Sufis see this for m of
dhikr as in feri or, this is because it is, we might say, easier than solitary
dhikr, which is not only practiced alone but also in silence and immobility.
The fact that music - or more precisel y the practice of music - occupi es a
central position in the first and is totally absent from the second, is not
fortuitous. If collective dhikr is easier than the solitary, this is because the
practitioner of collective dhikr is supported - literall y at times-by his
companions, whom he supports in turn. T his great ente rpri se of autosug -
gestion, namel y dhikr, is based on the synergic effect of mus ic. It is wi thin
music and through it; it is by means of the coordinating power that music
alo ne possesses that the fusion of all these wills and all these desires is re al-
ized, a fusion that prefigures for each one the communion with God. I
have made it plain enough by now: the singing and dancing practices in
dhikr should be seen as act ing in a very real way on the physiology of
trance . Their efficac y in the field of its socialization, in othe r words of its
normal and regular realization, is cenainly no less decisive . There is abso-
lutely no need, in order to explain their role, to lend them some undefined,
mysterious power.

Fakirism and Identificator y Music

In the last few pages we have concentrated on the orthodox form of dhikr,
if that is the correct way of putt ing it. But, as we have seen, there are oth-
ers, and it can happen that dhikr will include episode s much more akin to
possession than to Islam. Thi s is true , as we saw, of the skewer ordeal de-
scribed by Poche 1 i; among the Aleppo Rifa <iyya, during which the neo-
phytes, whose flanks are pierced , identify themselves with Ahmad Rifa ' i ,
the orde r 's founder. During this episode it is not the chanted recitation of
the dhikr that is heard , but a mixture of songs and drumming, 1 1 ' in which
the principal drum instrument is the mizhar, whose name means "he w ho
makes appear." According to Poche, this drum plays an essential role in
the neoph ytes ' entry into trance. "The vigorous beating of the mizhar," he
writes, "gradually ove rcome s all physical resistance and ... produces a
sort of anesthesia that facilitates submission of the body to the discipline
of the ceremony, particularly during the sword ordeal. The loudness of
sound produced by the instrument, which can be heard from quite far
away, causes the accompanyi ng song to be more deeply impressed upon
the mind and body of the faith ful identifying themselves with him."
310 CHAPTER SEVEJ

Through his songs, the shaikh urges the candidate to prepare himself fc
the ordeal. In voca tion s to the "saint" mingle with the tireless beating c
the mizhar. Finally, "the entry of a rhythm in double time irresistibly urge
rhe murid to move to the centre of the halga [circle ]" where the ordeal i1
self is to take place. With respect to the relation between the music and th
entranced person, the situation is thus quite different here from that whic
characterizes the dhikr proper: it is not the musical practice of the dhikr b
the adept himself that leads him into tr ance , it is music that is pla yed fc
him. The practitioner is completel y musicared; he is absolutely not a ml
sicant. Moreo ver , the situation is also different from that of samii ' (wher
aga in the adept is a musicated person, nor a musicant ), since it is not th
emotion - both aesthetic and mystic - experienced while hearing a rel.
gious song that triggers trance. This is done by songs urging him to be col
rageous, or exalting the saint, or else by codified drum calls, in short, ml
sic that has the same char acteristics as that found in the possession ritu al
described in the first part of this book.
The fact that the ordeal in question takes place during a ritual identifie
by the word dhikr does not alter the situ at ion: we are not dealing at a
wi th a communion rite , but w ith a possession rite , and wit hin the logic c
the system, it is possession music that must be at work.
ow let us move on to the ' Isa wiyya, whose dhikr, or rather~ acfra (ce1
emony ) includes aspects th at are even less orthodox than those of the Rifi
"yya. The seco nd part of this ~acfra consists of figurative dances and derr
onst rations of fakirism (ordeals of hot coals , daggers, and so on ) that mu:
all be viewed, as I have said, as manifest ations of possession. The music .
no longer provided (even in part ) by rhe adepts but onl y by mu sicians esp<
cially appointed to the task . One does nor hear recitation of the dhikr, cor
trar y to what happened during the first part of the ~acfra, but instrument,
music performed by drums and wind instruments-the oboe (gha itc
among the ' Isai yya of Morocco, 11 5 the tr ansverse flute (guec;ba)116 amon
those of Algeria. There is every re ason to believe rhar the figurati ve danct
and the demonstrations of fakirism (which are their equivalent ) are pei
formed to specific musical themes . Brunel ( 1926; 98 ) makes a very sign if
cant observation on this point . Talkir.g of a particular region where th
_hacfraof the Gnawi yya 111 is practiced, he notes that "upon hearing th
tune appropriate to this ~acfra, the adepts are seized by ha!. 1 18 The y ar
then offered daggers, which the y seize . .. [in order to perform ) their fa
vo rite exercise, wh ich consists of striking their calves with the points c
these weapons until the blood flows." othing is said abo ut the music;
characteristics of this tune, undoubtedl y because none is espec ially n
markable and because it does not significantl y differ from other tunes i
the repertoire , except for the fact that it is "approp riate. " We are justifie
in thinking, then , that this "appropriate tune" is nothing other than th
Music and T ranee among the Arabs 311
"right tune" to wh ich we have referred on several occasio n s in ot her chap-
ters , in other wo rds, it is the "agreed -up on" air, whic h has the value of a
signal and is recogn ized as such, but wh ich is not endowed wit h any par-
ticular expressive property peculia r to it . If the adepts go into tran ce w hen
they hear it, it is not because the tune has emotive powers specific to it, but
bec ause it is the tune of the Gnawa, and because at its cue one ado pts the
con ventional form of beha vior - th at is, one grabs the dagge rs. This "ap-
propriate tune" should therefore be placed in the category of musical mo t-
toes, and we ha ve alread y examined the mechanism by which the y act to
trigger possession trance. 119 This mechanism is namel y that of a code, in
which the signal triggers a response as a result of the meaning that has
been arbitraril y assigned to it, and not by some intrinsic power act ing
upon the auditor's emoti vity.
Let us now speci fy this interpretation of the role th at music pla ys in the
p art of the ' lsa w iyya f?acjrain which possession is at wo rk; it is one that I
deduce from Brunel's description an d is by no mean s his interpret ation of
the facts. Quite the contr ary. A "furious tempes t of drums and oboes" tha t
puts the dancers in such a state that "t hey can sca rcely retain their bal -
ance"; !20 "din"; I 2I "dizz ying music " th at , toge ther With the "demoniaca l
dance" of the tahhayor (generic term for tr an ce dances among the
'lsawiyya ), plunges everyone "into an acute neurop arhic scare"; 122 specta-
tors who dance "in the grip of the terrifying staccato music of the ghaita
and the thobol (oboe an d kettledrum ) and are seized by "t his contagious
madness" and "thi s ecstatic frenz y" whic h "none can resist "; 12 i "hyster ical
crisis" that abates "su ddenl y w ith the ghai"tas' last nore s""' - such are the
terms chat Brunel uses to describe how the music functions in this ritual.
Moreover, on several occasio ns he also mentioned the accele rat ion of tem-
po, so that rhe dances always beg in fairly slow ly and th en grad u ally speed
up . Clearly he pro vides us wi th a gener al picture of the rela tions between
music and tr ance that is extreme ly simila r to the one that emerged from so
many ocher ethnographical descrip tions cited in the first part of this book.
And in chis picture, as in all the ochers, it is imperative to distinguish be-
tween facts and interpretations of fac es. The facts are not to be doubted :
the music is extremely violent, it is markedly rh ythmic in character, accel -
erates, accompanies dances that have the same formal character istics, and
the danc er s are finally entr anced . Thus far one can argue with nothing:
these are the facts and they have been reco rded. On the other hand, as
soon as one ma kes this music and dance responsib le- and entirel y respon -
sible - for tra nc e, one is int erpr etin g the fac ts, and it is this int erpret ation I
in fact dispute.
Music and dance on the one hand, trance o n the other , go together.
Th ere is a rel ation of contemporaneity that is undeniable . Bur by m aking
this relation into one of causality, Brunel commits an error, since in the
312 CHAPTER SEV EN

same cultu re area, n amel y Morocco, m us ic an d da nces that are just as vio -
lent and m ak e jus t as mu ch use of accelerando are per form ed every da y
without necessa rily leading to tra nc e, possess ion , or demo nstr ati o ns of fa-
kirism. As for the manner in whi ch he arrived at thi s er ro r, it invo lves a
process tha t has two phases. First, he attr ibut es cert ain ch aracteri stic s to
the music that it does not ha ve, but that it seems to hav e beca use they are
characterist ics of the accompan ying tr ance. Second, hav ing m ad e thi s
confusion, he then considers tr ance to be an effect of the music pr ecisely
bec ause it has these simil arities. In short, the procedure is as follows: this
m an is mad; he dances and listens to music; therefore there is so methin g
m ad abou t thi s music; this being so, it is thi s music th at makes the man
m ad. Needless to say, this very system of interpretation that is proje ct ed
o nto the fac ts remains implicit. It is because one is watching tr an ces that
can indeed be terrifying that one labels the music as "terrifying" too. It is
beca use entranced people stagger about that o ne refer s to tr ance as "d izzy-
ing" and th at o ne speaks of th e "furious temp est unleashed by the drums. "
Ir is because the spirits possessing the ad ept s a re regarded as demons that
th e dance is la beled "demoni acal." In other contexts, the same m usic and
th e same d ance \Vould be described as "vio lent " certainly, but t hey would
no r be considered as either "terrif ying " o r "demoniacal."
As we know, the ahwach is the custom ary dance of a large nu m ber of
the Berber "tribe s" in the Moroccan High Atlas mountains.' 2 5 It is danced
o n the spot. Standing closely packed one aga inst the other, the dance rs
bend their knees in time w ith one another, acco mp anying each beat with a
sort of pounding of the grou nd and a back-and -forth sway ing of the bod y
very similar to the one Brun el describes in the rbbani tra n ce dance. It
makes use of m us ic th at h as an extremely m arked accelera ndo, also com -
parable to th at used in the "ecsta tic" d ance s of the 'Isawiyya desc ribed by
Brunel. Beate n for all the y are worth, the drum s (bendir ) also make a
"din" one could we ll term deafening, if not actu ally diz zying. From the
fo rm al point of view, then, the ahwach displ ay s all the featu res to which
Brunel a ttributes the triggering of trance . M o reove r, he also desc ribes the
trance dance of the ' Isaw iyya as a "dan ce of jubilation" (1 926; 122,
225 ). " 6 And an intense impres sio n of jub ilation is also what eme rges from
a su ccessful ahwach . Yet the fact rem a ins th at the ahwach is not a trance
dance at all. 12 '
Another argument, this time concerning the musical instru ments , fur -
ther streng then s my skepticism with regard co Brunel 's theor y about the
effects of the music. He att ributes the se effect s largel y to the oboe (ghaita );
yet it so happens th at the flute is sometimes u ed in its place. In h is descr ip-
tion of the ' Isawiyya's annu al feas t in M oroc co, Brunel (1926; 122 ) close -
ly associat es th e "hysteric al crisis" with the gha ita. The crisis, he writes,
aba tes "sud denl y with the final n otes ." He aiso emphasizes the ghaita' s
Music and Trance among'the Arabs 313
"nasal sound" (ibid., 98 ) and denounces its "ha tched an d frightening" mu-
sic, which, accompanied by the drums, leads to "frenzied tempests." As
we know, the ghaita, the oboe of the Maghreb, is an in stru ment with a
piercing, very intense sound. Clearly, it .is its piercing quality and its tim-
bre sui generis that Brunel interprets as being responsible for triggering
trance. 128 Yet among the Algerian <Isawiyya it is not the ghaita that is used
for the ~acfra and trance dances bur rhe transverse flute, an instrument
whose sound is always rather weak .
In their description of trance dances amo ng rhe Algerian <Isaw iyya,
Dermenghem and Bar bes ( 19 51; 309 ) talk about the "irresistib le beating"
of the bendafr, and thus attribute rhe same effects to the drums as Brunel
does. As for the flute (guer;ba ), it accompanies the drums "d iscreetly," they
rel! us, and irs song is "so gentle that ir can pass unperceived." The y also
add, however, that this song is "so pure that the nostalgic appeal of its
melody colors the imperious rh ythm of the percussion instruments wirh
great efficaciry." What sort of efficaciry would chis be? In Morocco, ir was
the violence of rhe oboe that combined w ith rhe drums to trigger trance; in
Algeria ir is rhe gentleness of rhe flure. Some explanation seems to be
called for. The truth is rhar both these interpretations are equally subjec -
nve.
We are once again confronted with the findings in rhe first part of rhe
book on rhe nonspecificity of the musical instruments associated with pos -
session. 12 9 In practice, as we saw, rhey all can serve chis purpose. This ex -
ample further confirms this .
I should add char in Morocco, among the Gnawa, anothe r brotherhood
rhar has numerous "points of resemblance" (Brunel 1926; 234 ) with char
of rhe <Jsawiyya, ir is nor even a wind instrument, but a stringed o ne, the
ge mbri, a small lure, chat is used as the melodic instrument in the band ac -
comp a nying trance dancers. Here aga in, however, rhe gembri, which is
"the fundamental instrument " (Lapassade 197 5; 30 n. 3 ) of the orchestra,
and \\·hich produces only rather weak sounds, is "generally" covered
(ibid., n. 5 ) by rhe "deafening" music of metal castanets and drums. How
do we explain the fact th at an instrument can play a "funda ment al" role in
a possession ' 30 ritual and yet be scarcely heard? Clearly because the func-
tion of the gembri (like that of the oboe and flute earlier ) is to announce
the musical mottoes of the genii responsible for the possessions, and be -
cause all that is needed in order to produce the desired effect is for the
adept consecrated to a particular genius to recogni ze that spirit's rune
(which is easily conceivable, however great the surrounding din ). This is
the onl y hypothesis through which this paradox can be solved.
Even though the musical motto, or in other wo rds the melodic asp ect of
Gnawa music, plays the essential role in triggering trance , it is ne verthe -
less only its rhythmic aspect - whi ch is much more spectacula r-th at
314 CHAPTER SEVEN

caught the attention of the Living Theatre ' some years ago. It seems that
30

it was also this aspect that attract ed a yo ung Moroccan, Pacca Abderha-
man, founder of a pop group whose aim is also to induce trance through
music, but in a pro fane context. ' 3 ' And this is further proof that such erro-
neous ide as concerning the preeminence of rhythm in the relation between
music and trance reign everywhere. To repeat once again: these ideas are
the reflection of a much too simplistic view of things .
8 Con clusion

If it has taken so many pages to make this inventory of the rela ti ons be-
tween music and trance, it is because the y are so extremely varied, becaus e
the y often contradict one another from one case to another, and because it
is extremely difficult to formulate any rule about them w ithout a counter -
examp ·le immediately contrad icting it. At the same time, things had to be
shown in all their complexity so as not to fall pre y to interpretations of the
most reductive and erroneous sort. At least one fact has been established:
that these relat ions are not simple. Although the y are complex, this does
not mean that the y defy all logic. And elucidat ing this logic is the task that
now rema ins. The only way to do this is to seriate the questions, which we
shall do by successively examining all the principal types of trance as they
have been defined in the previous pages, starting wi th their ideology, and
by relating these trances to the manner in which tbey are engendered (in-
duced / condu cted trance ), the dynamics of their m anif esta ti ons, the kind
of mu sic associa ted with them , and the role pla yed by dance . Among all
the se types of tr ance , possession trance is the one that seems to ha ve the
most paradoxical relations to music. I shall therefore examine it last, and
at greater length.

MUSIC IN ITS REL ATIONS T O E MO TI ONAL , CO MMUNIAL,

A. ·o SHAMAN IC TRA NCES

Among all the types of tr ance we have discussed, emotional tr ance-pro -

fane o~ religious - as we observed it among the Arabs is und oubtedly the
one th at has the most direct and evident relation to music . Upon hearing
music that has a strong emotion al power over him, the subject, over -
whelmed by emotion, goes into trance . From w h at source does the music
draw this power? From the meaning of the words and from the perfection
of their rel at ion to the music. The emotion is not only affective; it is to an
equal extent aesthetic, and calls upon either the subject 's sense of th e beau -
tiful, o r his sense of the divine , or upon both simultaneously. In this case
music is pure message , and it in fact produces trance by virtue of a sui gen -
eris p owe r. It must nevertheless be emphasized th at this power is ind isso-
ciable from that of the wo rds - it is the "union of poetr y and music ," as
316 CHAPTER EIGHT

they said in the Renaissance, th at is at work. The relations between music

and trance are thus, in this instance , as simple as they can be. T he only
problem the y present is that of knowing why musical emotion so com~
manly leads to such conduct among the Arabs and does so less frequently
elsewhere. The answer is clearly that this behavior among the Arabs is a
cultural phenomenon, an expression of emotions that have been learned,
stereotyped, and assigned a certain va lue. And although the phenomenon
is less usual elsewhere, we do have some examples of it. In Europe, I cited
the "Ranz des vaches" and its effect on the Swiss. We have also recentl y
learned that in Melanesia, among certain New Guinea Papuans and also
amo ng the 'Are 'a re of the Solomon Islands ,' musical emotion frequently
gives rise to certain forms o f trance. In the West, the emotional impact of
military bugle calls, which can reduce the bravest men to tears, is another
example. There is also justification for thinking that the more or less hys -
terical behavior of pop music fans can be categorized as musical emotion
that has become an accepted and va lued stereotyped form of behavior.
Of all the arts, music is undoubtedly the one th at has the greatest capac-
ity to move us, and the emotion it arouses can reach overwhelming pro-
portions. Since trance is clearly an emotional form of behavior, it is not
surprising that musical emotions should prove to be destined to some ex -
tent, under the pressure of cultural factors, to become institutionalized in
this form. This would mean that we are dealing here with a relation be-
tween music and trance th at, although strongly influenced by culture, is
nevertheless based on a natural-and thus uni versal - property of music,
or at least of a certain kind of music. Moreover, in emotional trance it is
mu sic alone that produces the trance. This trance, wh ich frequently is
nearer to nervous crisis than to true trance , is of short duration, since it
does not result, or only to a very small extent, in dance. Contrar y to what
happens in the case of ocher sorts of tr ance, music does not function to
maintain it.
Communial trance, as I have said, occurs in rwo forms, induced and
conducted . When induced - meaning th at the subject is a music ated per-
son and not a musicant - it is always emotional, bur can be either ritual-
ized or not. When it is ritualized and blossoms into dance, it is the classical
Arab sam?i< . Music brings about the trance as a re sul t of the emotional
power - entirely steeped in religious feeling in this case, to be sure - of the
sung poetry; it th en prolongs it- maint ains it, we could say-during the
dancing by means of its rh ythm. Ir therefore has rwo effects: first, trigger-
ing the trance; second, maintaining it-the two resulting from two very
distinct actions, first that of the words, then that of the rhythm.
When communial tr ance is conducted, in othe r words when the subject
is a musicant, we have the Muslim dhikr or, among Christians, the tranct
practices of the Shlustes, the Shakers, and various other sects. Here , tht
Conclusion 317
role of the music in producing tr ance is of a quite different kind. Being in-
dissociable from dance - for it can only be made w hile dancing-it is
above all a corporeal technique. The dancer and the musician, or more
precisel y the dancer and singer, are merged in one and the same person. It
is both the subject's own singing and dancing that lead him to trance. The
dhikr is a certain way of singing. But th is singing, w hich calls for a par-
ticular technique of breath control, is at the same time a recitation of the
name of God (among Christians it refers to the Hol y Ghost or Ch ri st ). Th e
words also play a role (and many would say that it is the essential one ) and
are, here again, bearers of emotion - religious emotion, naturally, but ar
the service of a religion of love and fervor conducive to man y amb iguit ies.
In content and in form chis music is above all invocarory. If one dissociates
it from the meaning of the words, this music - sinbing and dancing com -
bined - seems ro have rhe function of creating excirarion. What we are
dealing with then is a very particular form of autoexc irar ion, since ir
makes use of breathing, a certain oversi imul ation of the vocal cords, a
very accentuated rotary movement of rhe neck and head, and a whole va-
riety of physical movements char muse certainly consume (or liberate? ) a
great deal of energy. There are rw o musical characteristics specific to
dhikr : a particular sort of vo cal delivery (the "saw" dhikr )- which raises
rhe question of whethe r or not it has neurophysiological repercuss ions -
and a systematic use of accelerando and crescen'do , the object of wh ich is
ro increase the emotional tensi on. Although it is not as monotonous as
people rend ro claim - even if only because the increasing agitat ion just
mentioned contributes to a renew al of the musical interest -r epetit ion
(and above all the repetition of wo rds ) also pla ys a role . Let us say, to sum
up, rhar music, words, and dance create at rhe same time a great physical
effervescence and a scare of "mono ideism " rh ar, in combination, create
psychophysiological conditions appare ntl y very favorable to the occur -
rence of trance . In this case tr ance is quite deliberately sought for, and if it
occurs - which is not always the case, for rhe "mechanism " is not auto -
matic, eve n though ir is certainly very efficacious - it is only ve ry rarel y
char ir does so in an unexpected_ manner, unlike emotional trance , wh ich
very often does occur unexpectedly.
In the case of both induced communial trance (sama<) and conducted
communial trance (dhikr and other comparable rituals ) the onset of trance
requires a certain amou nt of theatricality. But the theater involved is as it
were an intimate - or innermost-one. Fo r al th ough there may be an au -
dience (varying in size depending on rhe case, and some time s completely
absent ), the presence of specrarors is not indispensable. Here the theater is
constituted by the small group of adepts raking part together in the prac -
tice of either "aud iti on " or sung recitation and dancing. The kind of dhikr
practiced aloud - in othe r words the one that interests us here, since it is
318 CHAPTER EIGHl

the one that leads to trance, whe reas the other leads to ecstas y-is termec
"collec tive ." Samci' is likewise a group pr act ice, not only because there ar,
always several people taki ng p art in it together (I kno w of no account ~
samci' organized around a single indi vidu al ), but also because the gat her
ing is necessaril y divided into at least two sect ions, "a udirioner s" on th ,
one hand, musicians on the other.
Th e Mevlevi samci' , as I have said, is a particular case in that it is akii
both to classical samci' and at the same rime to dhikr, since the dancers ar
not the mu sicia ns (or the musicants ) of their own entry into trance , where
as the ir dancing is the principal me ans of triggering it . Contrary to w ha
happens in classical samci' , dance is not the result and expres sion o
tr ance; rather, trance is the result of dance . The music here certainl y i
emotional since bo th words and instruments are impregnated with love c
God. But it is the dance, ultimatel y, that is the tri ggerin g factor. The rol
of the music thus is twofo ld. First, it creates a general state of myst ic em c
tion -t he emotion at once religious and carnal that is so char acteristic c
Sufism - and second, it provides the dance with the acoustic stimulu
without w hich it could not even take place. Let us say, to sum up , that her
the role of music in the triggering and maintenance of trance is ind irec
rather than direct, contrar y ro what happens in both classical samci' an
dhikr.
No w let us move on to shama nism, in which the trance, as in dhikr, i
con ducted , since the shaman is the musicant of his own entry into crane
(the fact that he is aided in his role by an assistant w ho takes over when th
moment arises, notabl y when he loses consciousness, does not inva lidat
this rule ). T he relations betwee n mus ic and tr ance are organ ized , her
again, in acco rd ance wi th a particular system, but matters are more corr
plicated than the y ha ve been up to no w. T he shaman's musical instrumer
is endowed wi th symbolic meaning s related to his journe y, or, more pn
cisely, to the world or worl ds he visits during his trance. If the drum-sur
posing the instrument is a drum, for even in central Asia, remember , it ca
be some ocher instrument-p lays a role in triggering trance , this is not, d1
spite w hat ma y ha ve been said to the contrar y, the result of some myster
ous neuroph ysiological action spec ific to th at instrument, as I have show
onl y roo often, nor of some kind of "o bsessi ve" monoto ny that also exis·
only in the imagination of certain authors. Musicall y, the shaman
drum - or an y other instru ment used in its place-essentiall y functions t
support his singing, to pro vide the rh ythm that is the prim ary support <
his dancing , an d to dramat ize or punctuate the act ion. In short , its role
prec isely the same as the one it pla ys in theater music of an y kind, w ith d
sing le exception that here it is ch arged with symbolic meaning , and th ;
this symbolic meaning is in turn ch arged w ith a certain emotional powe
But even in this case we are still in the realm of ps ycholog y and cultur
conditioning.
Conclusion 319
As for the shaman's singing, it has several different aspects; sometimes it
is invocatory, when he has to summon his auxiliary spirits, someti mes de-
scriptive or narrative, when he is relating his journey, but its specific char-
acteristic is that it is incantatory . The shaman is a magician, and his sing-
ing brings to life the imaginary world of the invisible. Without song, the
shamanic im aginary system would be inconceivable. (The imaginary sys-
tem of possession, on the contrary, can dispense with it and make do with
purely instrumental music, which operates by means of coded signals
whose verbal equivalence is secondary .) Moreover, for the shaman, sing -
ing is his principal me ans of communion with his audience, that audie nce
without which he could nor perform and wh ich supports him with its cho-
ral responses, for it is the alternation of his calls and their replies that cre-
ates the reciprocal warm -up, the climate of emotional excitation, that is
indispensable to the onset of the trance. Lastl y, for the shaman, to sing
and play the drum (or rarcle ) is to srim4late himself to dance. Beneath the
often considerable weight of his costume, this dance rakes on the propor -
tions of an exhausting exercise, responsible, beyond doubt, for certain
losses of consciousness. To shamanize, in other words to sing and d ance,
is as much a corporeal technique as a spiritual exerc ise. Insofar as he is at
the same time singer, instrumentalist, and dancer, the shaman, among all
practitioners of trance, should be seen as the one who by far makes the
most co m piece use of music.
But che power of music alone cannot be held re sponsible for the sha -
man's entry into trance, anymore than in the case of the Sufi. This trance
muse still be willed. Let us remember what Shirokogoroff observed on chis
point : the accelerando and crescendo of the music , intended to intensify
the general emotion , do not lead to trance unless the shaman has decided
that the y will at the outset. Neither the music nor the dance produces
tra nce mechanically or automatically, and chis explains why the sha -
man - just like the Sufi- so often resorts to the use of a drug in order to
obtain "lift-off," as it were, at the start of his journey. W ith a drug, the
mechanism is different and more reliable. Music and dance on their own ·
are certainly the instruments of a more authentic, and more meritorious,
trance; bur they also are much more chancy.
W ith the shaman we are dealing with someone whose trance - or, more
precisely, nervous crisis or fir- is not necessaril y linked to music. Every
shaman, before becoming a shaman, has undergone sudden and unfore -
seen crises chat occurred outside of any ritual context, and that were in
face the very sign of his vocation . Th e crises, or trances, that he w ill regu -
larly experience after having been officially established as a shaman, will
be no more, in part at least, than reenactments of previous crises. And it
must be emphasized that the crisis - that particularly dramatic aspec t of
trance-does nor require music to trigger it. The psychic upheaval that it
manifests thus obeys a purely internal logic of the stare of consciousness.
320 CHA PTER EIGH T

Thi s means that the role of the music is much less to produce the tr ance
th an to create conditions favorable to its onset, to regularize its form, and
to ensure th at inste ad of being a merel y indi vidu al, unpredictab le, and un- ·
controllable behavior al phenomenon, it becomes , on the co ntr ary, pre-
dictable, controlled, and at the service of rhe group.
As o bser ved amo ng the Bush men , tr ance h as yet anothe r set of relations
to music. As in sham ani sm, to w hich it is kindred in at least one aspect of
its ideology (journey of rhe soul a nd ascent into the upper wo rld ), we are
dealing with a conducted tr an ce in w hich the subject is the musicant of
both his own d ance and his ow n tr a nce . As in communial trance (as mani-
fested in dhikr, for example ), it is a collecti ve practice; but the singing,
wh ich is the m ain support for th e dance and has no instrument al accom -
paniment ocher than the dancers' own leg-rattle s, is d ivided between the
men wh o dance and wi ll go into trance and the wo men who do nor dance,
are sea red , provide the greater p art of the cho ral singing , and do n or go
into tr an ce. T he singing, wh ich is symbolic ally associa ted w ith one or an-
o ther healing ritual , is wo rdle ss . It is vo calized on syllables devo id of
meaning and does nor constitute a p articul ar mu sical category. The same
type of sing in g th at is usually intoned in everyday life is used for the he al-
ing dance, w hich gives rise tO tr ance. Thi s singing is neither invocat ory
nor incantatory, and it is not its emotional power char is at work, but the
ardor of its collective execution, because of w hich the cur ative power of
rhe sing ing "boils" and thus att ains its full efficacy. The psychological ef-
fervescence, the emotion neces sary for th e onse t of trance, is thus deri ved
from the collect ive fervor of the singing . As for the dancing, w hich like
char of the shaman or that of the dhikr , th oug h for diffe rent reasons, is an
exhausting exerc ise, it also contributes ro th e ph ysiological preparati on of
tr ance . H ere agai n , music an d dance , taken rogether, const itute a cor -
poreal technique aim ed at atta in ing tr ance.
"U nderl ying all our mystic states," Mauss w rites (1936 ),' there are
"corporeal te chn iques ," "biolog ical meth ods of entering into communica -
tion wi th God." If we substitute the wor ds "re ligious trance" for "m ystic
states" and "commun ication with God " tO make our gene ralization clearer, it
is evident that for the Bushmen of so uth ern Afric a, as for the shamans of
central Asia, the Russian Shlustes, the Sufis of the ear East , or the Shak -
ers in the Uni ted States , tr ance is very largely a matter of corporeal tech -
nique , in which singing and dancing, in combi n at ion, ar e the two princi -
p al elements . But ir would be a fun d amental error to reduc e these va rio us
fo rm s o f tr ance ro no more than va rious forms of corporeal technique us-
in g va rious combinations of song and danc e. Th e techn ique operates only
because it is ar the ser vice of a belief , and because trance co nstitute s a cul -
tur al model integrated inro a certa in general representation of the wo rld .
H ere we h ave an essenti al intellectual datum, wh ich underlies both the
Conclusion 321
psychology and ph ysiolog y of trance. This is why entry into trance always
seems to depend upon a kind of restrictive clause: however well prepared
one may be, physicall y and psychologicall y, one must still be prepared in-
tellectually, and have made the decision (more or less unconsciously ) to
succumb to the trance state.
The great difference between emotional trance on the one hand and
communial (conducted ) and shamanic trance on the other, is that in the
former the sole responsibility for the onset of trance rests upon listening to
music-or, more precisely, to sung poetry-while in the case of the other
two it is not listening to music but its practice that is invo lved; furthe r-
more, it is not the practice of music alone that is the operative factor, but
this practice in combination with dance. In the first case the music is in no
way a corporeal technique; in the other two it is fundamentally and pre-
cisely that. Moreover, in the first case the music is essentially emotional; in
the other two, whether invocatory (communial trance ) or incantator y
(shamanic trance ), it is much less emotional, or onl y indirectly so. In the
case of the Bushmen it seems it is not emotional at all, which rai ses a prob-
lem I shall leave open.
Things look quite different, however, in the case of possession.

M USIC IN ITS RELATIONS TO POS SESSION TRANCE

The reader will remember that in the first part of this book, we found that
the rel ations between music and possession can assume an extraordinary
number of different forms . Ar the moment of the actual triggering of
trance, which is often but not always marked by a crisis, it took on the
most contradictor y aspects. The crisis sometimes appears robe the direct
result of the music's frenzy, sometimes quite unconnected w ith it; music
sometimes appears to be the means by which the crisis and fall (in the
ndop) are pro voked, some times, on the contrar y, as the means by which
the crisis is brought to an end (as in ra rantism ); sometimes it appears to be
an almost mechanical means by which trance is pro voked (in public can -
do m ble ceremonies ), sometimes it seems to play no part in it at all (the
morning trance during the can dom ble re clusion period ); the adepts some-
times appear to be incapable of resisting the summons of the music (when
they are neophytes ), sometimes, on the contrary, it is a point of honor not
to succumb to it (when the y are off iciants ); in one ceremony an adept goes
into tr ance at the sound of his motto, while another, apparently subject to
the same objective conditions, does not; in one cult (candombl e) each god
responds onl y to the musical theme (motto ) specific to him, yet there also
is a theme that can mobilize all the gods indiscriminatel y: one dance (in
the zZir cult ) is reputedl y a class ic means of inducing trance, yet is also per -
formed to provoke the coming our of trance; in the rab cult, the drummers
322 CHAPTEREIGHT

appea r to be pla ying a determining role in the onset of possession when

great public ceremonies are involved (tuuru and ndop), yet pla y none
wha tsoe ver w hen the ritual is on a smaller scale (samp ) since they are
strictly excluded; sometimes the instruments are held responsible for the
onse t of trance, sometimes the singing; here it is one type of instrument
that is characteristic of possession music, there it is another. In other
words, although music indisputably plays a role-more or less decisive
depending on the case-in the onset of trance, it seems impossible roes-
tablish an y constant relation of cause and effect between the two. The
only rule that appears to be trul y general is that the music must be made
for the possessed and not by them, or, in other words, that trance is in-
duced and not conducted .
If we now go on ro examine trance, not at its moment of onset, but
whe n it is at its height, things become clearer. We could sum them up by
sayi ng that no matter wh ich cult is concerned-if we set aside the initi-
ation period which has its own particular logic - possession music is
dance music whose melody is associated w ith the divinity held responsible
for the trance. The relations between music and trance constitute them-
selves in two different ways depending on whether trance is at its moment
o f onset or, on the contrary , when it is in full bloom. The y are also of two
different kinds depending on w hether the possessed person is at the initi-
ation stage or not. In order ro understand the reason for this state of things
we must now look at the relations between music and possession with re-
spect ro the logic that governs the system since, indeed, a system seems ro
be at work . In order ro do this, however, it is indispensable robe able to
refer ro a general theory of possession trance. To my knowledge, no such
theory exists. 3 However presumptuous it may seem, I will therefore run
the risk of offering one here, less for its own sake (even though I believe it
does account fairly well for a number of facts ) than because one is needed
anyhow . Good or bad, it will have the merit of occupying, until something
better comes along, a position that another theor y of this kind is bound to
hold in thi s system.
Provisoril y, then , let us view trance, as it is observed in possession cults,
as a so cialized form of beha vior resultin g from the conjunction of several
constituents: (1 ) at the level of the indi vidua'.: a given inn ate ro the struc -
ture of consciousness making it susceptible of being invaded by an emo -
tional event that submerges its normal state and leads ro hysteriform be-
havio r ; (2 ) at the level of collective representations: (a ) interpretati o n of
this event as a sign of the will or presence of a spir it or divinity; (b) exploi-
tation or, to use Roger Bastide's term (1972, 94 ), domestication of the
event, with the intention of establishing it as a mode of communication
with the divine; (c) identification of the entranced subject wit h the divinity
held resp onsible for the trance; and (d ) theatrilization of this identifica-
ror y behavior.
Conclusion 323
T he way in w h ich these constit uent s are brou ght in to pl ay varie s a great
deal from cult to cult, de pending on the cul t's history, the circum st ances in
wh ich it was gr adu ally co n stitu ted, the cont ex t in wh ich it developed , the
ind ivid u als w ho aff ect ed its evol uti on , its poss ible co nt acts w ith nei ghbor-
ing or foreign cu lt s, and still a great many ot her th ings. For m any adept s
the origi n al em ot ion tha t gave rise to the m ent al uph eava l aro und which
the cult struct ur ed itself may well not or no longer be a spontaneo u s form
of behavior but, on the contrary, a fo rm of learned behavior, a ste reotype.
This does not alte r the fact that in the genesis of all possession cults, and
therefore all possessional trance, the re mus e, in the begi nning, have bee n
an emotional state experienced by an individual and capable of being ex -
perienced in turn by others, whether spontaneously or as the · result of a
learning process. The relat ion of mus ic to this emotional state is extre mely
va riable since it is integrated into the cult in varying ways. Initiation, ac-
cording to whether it is very important (in the vodun cult, for example ) or,
on the contrary, nonexistent (as in tarantism ) go verns the emotional be-
havior and its relat ion to music sometimes in one way, sometimes in an-
other. Although emotion is something experienced by individuals, in th is
instance it manifests itself only within the logic of the cult into which it has
been integrated. This is what explains the disconcerting variability of the
relation bet w een music and the onset of trance.
The trance itself, in ocher words the period during which the sub ject set -
tles himself, so to sa y, into his ocher persona and totall y coincides wi th it ,
has, on the contrar y, a quite stable relation to music. In its fullest fo rm, by
which I mean the form it displays during public possession cerem oni es,
trance most often cons ists in dancing to the sound of music that belongs to
the possessing divinity or, if there are several persons, to the sound of m u-
sic that corresponds collectivel y to all the div inities present. H ere the func -
tion of the music is obvious. It is due to t he music, and because he is sup -
ported by the music, that the possessed person publicl y lives out, by
means of dance, his identification with the divinit y he embodies. T he mu-
sic at this point is thus neither emotional , nor invocatory, nor incantator y :
it is essentially identificatory. By pla ying his "motto," the musicians not ify
this ident ity to the entranced dancer, those around him, the priests, and
the spectators. T he language the music speaks is understood by all, and
each person decodes it at his or her own level. It is through this music, and
through the dance to which it gives rise, that recognition of the div inity's
presence is conveyed to the entire group, a recognition that is indispensi -
ble because it authent icates the trance and confers upon it a charac ter of
normality . Music thus appears as the princ ipal means of socializing
trance.
The onset of trance very frequentl y obeys this logic as well, of course . It
is upon hearing his tune, his motto (or rather that of his god ) tha t the pos -
sessed person most often enters into trance. But the exceptions to the ru le
324 CHAPTER EIGHT

are so numerous that it is impo ssible to ignore them . Except in the case of
"old hands," going into trance constitutes, for the individual concerned, ,
such an event -I would even say such a psychoph ysiolog ical adve nture-
th at in order to make the transition successfully, a large number of favor -
able conditions must be present. The ind ividual equat ion - to be or not to
be disposed to enter into trance - then plays a decisive role in the process,
and this is what explains why music, which also has a part in it, indisput-
ably, but w hich is no more than one element among others , should ulti-
matel y pla y so very var iable a role. Sometimes it is its identificator y char-
acter that is at work, and then we are in the realm of cultural convention.
Sometimes it is its power to create, through its ability to excite and agita te
(hence the frequent use of accelerando and crescendo ) a certain state of ef-
fervescence that seems to be particularl y propitious, very underst andabl y,
to the onset of trance. And in thi s case we are in the realm of the natural.
Sometimes it is its associatio n wi th a particular circumstance, or idea, or
character, charged, for the adept , with emotional power. And then we are
in che realm of che individual.
In che triggering of trance , mus.ic contributes sometimes through one of
ics aspects - most often , again, its identificatory aspect - sometimes
through another, sometimes decisively, sometimes accessorily. In extreme
cases it ma y not contribute to it at all. This is true of tarantism, where, it
will be remembered, the tr ance-or, more correctly, the crisis (and fall )-
usually occurs outside any relation with music. This is an extremel y sig-
nificant fact. On the other hand, the tarantulee, whe n no longer in crisis
but in trance, dances to the sound of the tarantella, and it is in fact this tar-
antella (as long as it is th e "right tune" ) that ends the crisis and causes it to
develop into trance. Although it is conceivable that a subject can enter
into trance without music, it is inconce ivable that a subject could experi-
ence che trance itself without music. Let us say that, in possession, music is
the condition sine qua non of the trance experience. T his is so for two rea-
sons . First, because possession trance is a change of identity, because that
change of identity has no meaning for the subject unless his new identity is
recognized by everyone, and because it is the music that signals it. Second,
because chis new identity muse be manifested and because dance is (usually
with costume, but not always ) its principal and frequently sole expres-
sion. Provided , then, chat ic is not absolutely fleet ing (I am thinking of
Malkam Ayyah u's trances, described bv Leiris, which often lasted no
more than an instant, just long enough to express it with a gesture, word,
pose ), provided th at it has duration, this tr ance, which is the exper ience of
another identit y, has an absolute need for music in order to continue to ex-
ist, since it is mu sic chat, through its identificatory character, maintains
the illusion and that , through its function as support of the dance, enables
it to be m anifested.
Conclusion 325
The major functio n of music thu s seems to be maintaining the trance,
rather in the way an electr ic current w ill maintain the vibr ation of a tuning
fo rk if t une d to the sa me pitch frequency. Here, however, mu sic is not just
physically (on a pur ely motor level ) "in tune " w ith tr ance. It is even more
"in tune" on the psycholog ical level, sin ce its action consists in putdng the
individual expe rienc ing his tr an sito ry identity "in phase" with the group
that is recognizing this identity or impos ing it upon him.
Whereas in conduc ted trance-shamanism and even more so commun -
ial trance of the dhikr type -mu sic appears as a corporeal technique, here
it is above all a techn ique of communication. Since shamanic trance in -
volves changing wo rlds, the shaman's adventure is first and foremost an
ind ividu al affa ir. It cou ld even be practiced in solitude . And this is why the
shaman is of necessity the musicant of his own tr ance.
Possession tr ance, on the contrary, consists in a change of identity, and
this change would of course be meaningless if it were not recognized by
the group . Moreover it is, or is supposed to be, undergone , not willed .
This is why the music is provided by the group, since the possessed perso n
only exists as such for the group and because of it. In this relation , th en,
the music is the instrument of communication between subject and group;
bur the communication that takes place is particularly character ist ic in
that the two interlocutors speak simultaneously - one through music, the
other through dance - and that the dialogue is addressed to a th ird party,
the spectators. This third party is just as indispensable to the blossoming
of trance as the two others, since possession cannot function w ithout be -
ing theater.
Music has often been thought of as endowed with the mysterious power
of triggering possess ion, and the musicians of possession as the withhold -
ers of some mysterious knowledge that enables them to manipula te this
power . There is no truth whatsoever in this assumption. It is nonetheless
true that possession cults, as institutions, are mechanisms that make use
of great musical skill that h as developed over a long period of time. The
role of this music is multiple. At the level of the ceremony - or, if one pre-
fers, of the theater - it creates a certain emotional climate for the adepts.
Second, it leads the adep ts toward that great mutation, occuring at the
level o f im ag ination, that consists in becoming ident ified with the spirit
possessing him. In this operation, which is so aleatory and which is sub -
ject to so many var iables, the relation between music and trance often ap -
pears quire strained . Third an d last, it pro vides the adept with the means
of manifesting this identification and thus of exteriorizing his tr anc e. It is
at this stage that music is indispensable. Wh y? Because it is th e only lan-
guage chat speaks sim ultane ously, if I may so put it, to the head and the
legs; because it is through music that the group provides the entranced
pers on w ith a mirror in which he can read the image of his borrowed iden -
 326 CHAP TER EIGHT

tity; and because it is the music that enables him to reflect this identity
back again to the group in the form of dance. There is no mystery to i~ at
all. Or, if there is, then it lies in the trance state itself, as a special state of
i'
,. consciousness; and if we must seek for an explanation of this, it may be
i found in the overriding power of a certain conjunct ion of emotion and
l
'' imagination. This is the source from which trance springs. Music does
nothing more than socialize it, and enable it to attain its full development.

·.:

''-

c:
;.
 -Notes

C HAPTER O NE
1. On the universaliry of trance, see Erika Bourguignon 1973 , 9-1 1.
2 . Taber 's Cyclopedic M edical Dictionary, 9th ed, 1963 .
3 . Dictionnaire des termes techniques de med ecine, 16th ed ., Maloine, 1955.
4 . Cf. , for example, the captions to photog raphs 13a and 13b.
5 . The Book of the Right Usages of Audition and Tranc e. We will deal with this work ar
lengrh in chapter 7, "Music and T ranee amo ng the Arabs" in pan rwo.
6 . Speaking of srares of hype rexcirarion of the central nervous system, Bourguignon notes
( 197 3, 5 ) rhar trance arra ins "an exrreme in the ecsraric srare of mystical rapture." _Sheis here
following her own definition of ecsrasy in a pre vious work ( 1965, 41 ). Borrowed from the
dictionary of psychological and psychoanalytic rerms by English and Engli sh ( 1958 ), this
definition is: "Religious and emotionall y marked trances are called ecsrasy."
7. See pp . 182- 83, among orhe rs .
8 . In the Dictionnaire des religions (Larousse 1966 ), only rhe term "exra se" figures, pre -
sented as (mainl y for oracles ) rhe result of trance. In the Dict ionnair e de spiritualite, which is
in rhe process of being published and has nor yet gone beyond rhe lerrer L, a great deal of
space is allo ted ro rhe wo rd "ex rase"; various art icles, totaling 140 columns in all, appear un-
der rhis entr y. The word "rranse " oc curs several rimes under rhe entry on shamanism (1 961 ,
co ls. 2049 - 50 )- and this is significant.
9 . In his article on the definition of shama nism, Johan Reinh ard (1976, 16 ) tackles the
problem of how to distinguish between trance and ecstas y, bur wirhour arraching quire the
importance to it rhar I have here.
10 . Cf. "Vi e ecrire par elle-meme," chap. 20 in Oeu vres completes 1949, 193; Obras
comp/etas 1948 , 138.
11 . Ibid. 1949,308 .
12. My iralics.
13. Theres a of Avila 1949, 413.
14. Dictionnaire Bailly (189 4 ). Cf. the ind ex to Plato, Oeuvres completes, "Les Belles
Lerrres," vol. 13.
15 . On rhe meaning of ekstasis in Greek antiqu ity, see Dod ds 1951, 94-95.
16. See the word "rr anse" in Le Petit Robert dictionary.
17. On the ecstasies of rhe early Church Fathe rs, see J. Kirchmayer ( 1961 ).
18. Mireille Helffer (person al communic arion ).
19. And ras Zempleni is rhe one who told me of rhe existence o f these practices amo ng rhe
Wolof. T hey a re totally disrincr from rhe much berrer-known pr acrices of rhe ndop, w hich of
course involve trance. Evidentl y rhe khalwa is a borrowing from Sufism , and rhe word ,
mea ning "reclusion , rerrear," was used to form rhe name of the Khalwariy ya (cf. T rimingha m
1971 ).
20 . Thi s is Carder's suggesred rranslarion for sirr (195 2, 673 ). The public dhikr, which
leads nor to ecstasy bur to tr ance, will be discussed at lengt h in part rwo .
!!
I
328 NO TE S TO PAGES 8- 15

21. On nemb11tsu compared co othe r forms of mystic prayer, see G arder (195 2 and
1953 ).
22. In his study on the n atu re of yoga , printed as an introduction to T herese Brosse' s .
work (1963, i- xviii ).
23. "En srasis" is a word invented by Mircea E liad e (1948, 93 ) co translate the Sanskrit sa-
madhi, whos e meanings he defines as ·'unio n, coraliry; absorp tion into, coral concent ration
of the spirit, conjunction ." In Le chamanisme (1968, 326 ) he makes a clear distinction be -
tween "ensrasis" and "ecstasy": for the shaman, "th e final aim always is ecstasy and the ec-
static journey of the soul into the va rious cosmic regions, whe reas the yoga seeks enstasis, the
ultimate concent ration of the spirit, and 'escape ' from the Cosmos."
24 . The quote is tr anslate d from the French since I did not have access eithe r to the Ge r-
man edition (1921 ) o r the English tr ansl a tion, which is sa id to be quire good.
25. The accounts they both left of their experienc es of possession have been publi shed .
Oesterreich quotes large extracts from rhem , take n either fro m the Histoir e des diables de
Loudun ( 1716 ) o r from the Bibliotheque diabol ique, published late in the las t cenrury by the
pupils of Charcor.
26. In a late r chapter dealing with exorcism, I shall rerurn co this reprobate form of
t rance.
27. I say "in principle ." On this subject see the reservations exp ressed by M. Leiris (1958 )
in hi chapter , "Cons ciousness o r Lack of Ir in Protagonists of Possession Seanc es. " Howev-
er. I should point our that these reservations apply solely co the zar cult as observed by Leiris
in Ethiopia.
28. In this case, we return - albeit by anothe r path - co the concepts of Fischer (1969 ) and
Ludwig ( 1968 ) to which Erika Bourguignon ( 1973, 5- 8 ) re fers her readers in her pre senta -
tion of "altered states of consciousness ." The first is based on a contin uous scale of neuro-
physiological facrors determining a gre ater o r lesser degree of awakeni ng of the cent ral ner-
vous system . T he second is based on a similar continuum, but one that is expressed in term s
of the inducement of these "altered states."
29. In her desc ription of these Shake rs, Jeannette J. Henne y, from whom I have taken the
data, makes no distinction between the rwo stares and calls them both "tr ance ." Yer there
are, as she he rself clearly indicates, two ve ry different states involved. This is a case - amon g
many others - in \\'hich the distinction l suggest making between trance and ecstasy could
I·
have been usefullv employed.
30 . In India (C. Malamoud, personal communication ), the person seized by the inspi ra-
tion of rhe Veda is called, in Vedic Sanskrit, the "t re mbling one ." Oddly, in his w ell-kn own
book, Th e Drums of Afflict ion (1968 ), in \\'hich trembling plays a central role and manife st-
ly constitutes a fo rm of trance beha,·io r, V. W. T urne r never o nce uses t he wo rd tr ance.
31. Signified / signifier: I am using here, of cou rse, the well-known distinction m ade by
Saussure between the C\,·o as pect s of the sign.
32. cf..-\ Patristic Greek Lexicon ( 1961, s.v. "energeia").
33. Michel de Ceneau ( 1970, 33,158,205 ).
34. Or . so metimes, fallaciously. Fo r it can happen not only that the tr anc e is feigned (a
subject to which we shall come later ), but also that the exr raordinarr powers attributed ro it
are merely illusionist's tricks. Th is explains why the famous "magicia n" Rob ert Houdin was
sent to Algeria, in 1856, ro beat the "m arabours " at th eir own game . (cf . Sergent 1952 , 125 -
44 and 185-86 ). Ther e are indisputabl y cases of fraud. H oweve r, this is me rely a marg inal
and insignificant aspect of trance.
35. More precisely, of "conve rsion " hysteria as it has been known since Freud. The
mechan ism of the conversion , it will be remembered, "consists in a trans position of a psychic
conflict and an attem pt ro resolve that conflict into somatic sympro ms, either motor (e .g. pa -
ralys is) o r sensorial (local anesthesia or pain, for instance. .) Symptoms of conversion a re
Notes to Pages 15-20 329
characte rized by their symbolic meaning: the y exp ress repressed representa tions in physical
form" (Laplanche and Ponta!is 1967 , s. v. "co nversion" ).
36. Cuc out of the M ental Dis orders Diagn ostic Manual of the APA in 1952, it was re-
placed by "conversion symptom," notes llza Veith ( 197 3, 6 ), w ho makes several inte resti ng
references to possession .
37 . D.S.M. III, Diagnostic Criteria Draft . , Ame r ican Psychiat ric Association. Pre-
publication 1.5 .78.
38. See the issue of Confrontations psychiatriques entitled Hyster ie ( 196 8).
39. In his essay entitled "No rm al er an ormal" ( 1956 ), reprint ed in 1970, p . 15.
40. In this passage, Bascide is conce rne d with possession rather rhan shamanism. It is still
true, however, that he is talking about trance in general.
41. In his study on sym bolic efficacy - which, as we know, is an interpretation of a sha -
manic text - Levi -Strauss ( 1958 , 222 ) rec alls th at Freud hi m self had alread y conside red th at
"desc ription of psychoses and neuroses in terms of their psychological srrucru re would one
day give way to a physiological or even biochemical conception of chem."
42. On the concept of "alte re d scares of consciousness," irs origin, and its applica tions,
see Bourguignon's introduction ( 197 3 ).
43 . T he same wor d is used in Bengal to den ore the famo us sect of inspired musicians who
call themsel ves baul ("m ad men" ). On the music of rhe baul, see the record made by George s
Luneau (discog raphy ).
44 . The word "representation" is of cou rse taken here , and in the general course of chis
book, ,,;ich the meaning arcribuced to it in sociology eve r sin ce Durkheim.
45. Eveline Loe-Falck 1963 (1977, 7).
46. All refe rences are co che second edition (1968 ) of his book on shamanism, not ro the
first edition ( 1951 ).
47. Fo r a very broad critical view of Eliade's position s on shamanism, see Eveline Loe-
Falck 1973, 1-2.
48. Used as though rhev were synonymous, as we have seen .
49. L. Delaby ( 1976, 130 ), according to Shirokogoroff. On the sha manic jou rney as re-
lated co char of the spi r its and ro the life of the hunte r, see Del a by I 977 .
50 . See C. MacDonald, "De quelques manifestations cham an iques a Palawan" ( 1973,
11-2 0 ) and va rious contributions co the issues whe re chis study has appeared .
51. To my knowledge, Dodds was the first to formu late clearly rhe distinction chat oug ht
to be made bef\,·een shamanis m and possession: "some writers ... use the terms 'shaman-
ism' and 'possession' as if they were synonymous. Bue the ch aracteris tic feature of shaman -
ism is not the entry of an ali en spi rit into the shaman; it is che liberation of the shaman's spi r-
it, which leaves his body and secs off a mancic jou rne y o r 'psych ic excursion.' Supernatural
beings may assist him, but his o"·n personalicv is the decisive element" (I 95 I, 88 n. 43 ).
52. On rhe general relation of the shaman co his "guardian spirit" and ·'auxiliary spirits,"
see Eliade's chapter (1968 ), "Acqui ring Shamanic Power s." For the Tung us in particular, see
L. Delabr 1976, 83ff .
5 3. The place of '·possession" in shamanism and its spec ific characteristics in chat context
ar e sec forth in va rious p assages of Eliade's book , notably pp . 23, 89, 38 , and 399 of the
1968 edition.
54. Presenred in "Possession et chamanisme" and developed in pares 3 and 4 of "La folie
des dieux er la raison des hommes," stud ies rh ar follow one another in Pourquoi l'epouser?
(1971 ).
55. For Eliade ( 1968 , 388 ), the impo rt anc e rhac che incorporation of spi r its has acquired
in Tungus shamanism is a re latively recent thing, arcribucable to foreign influence. Tungu s
melodies are said to bet ray a "Chin ese origin ," he notes (ibid, 200 ) citing J. Yasser (1926 ),
which "confi rms Shirokogoroff's rhesi s concerning rhe strong Sino -Lamaisc influen ces on
 330 ' OTES TO PAGES 20- 37

Tungus shamanism." Shirokogoroff's wo rk , wt must remember , concerns the Tungus of

easte rn Siberia and Manchuria.
56. "Shamanizing to the lower world is practised amo ng all Tu ngus groups ," Shirokogor-
off ( 1935, 306 ) observes at the end of his long description of a shamanic journe y to the lower
world.
57 . On the problem of the definit ion of shamanism an d the rela tive importance that
should be acco rded in that definition to the jou rne y of the soul as opposed to possession, see
Johan Reinhard's article (1976, 12-1 8) which includes a critical examina tion of Eliade's
ideas on this point. Wi th ou t going into detail, let me say that the views expounded in this ar -
ticle are somewhat at variance with my own, and that the reason lies partly in what I shall
have to say later about pos session .
58. The Gold are a subgroup of the Tungu so-M an chu family, whos e territory lies in east-
ern Siberia. Cf. Delaby 1976, 7 and 12.
59. Cf. Dela by 1976, 67.
60. Italicized in the original.
61. Described in th is way, the situa tion is not specific to the Tungus alone . It has broad
analogies with that among the Yakut. Cf. E . Lot -Falck (1970 ) and the dist inction made be-
tween "incorporation and possession," and between the mi:ini:irik "playthi ng of the spirits"
and the "tamer" shaman.
62. Johan Reinhard 1976, 16.
63 . For La urence Dela by (personal communication ) the difference lies above all in this
last dist inction.
64. In "Formes et transformations de la possession" and "Vers le chaman isme" (197 1,
255-76 ).
65. Ecstatic Religion ( 1971 ), photo 11 b.
66. See P. Verger (1957, 150-58 and photos 83- 86 ).
67. On the word wajd, seep . 258ff.
68. Segurola (1963 ) s.v. vodun .
69. Cf. P. de Felice 1947, 160).
70. On the importance of identificatory behavio r in H aitian voodoo, see Louis Mars
(1953, 218 ) and his inventory of the "types of identification in some cases of possession. "
i, 71 . On these two words see th e second part of this book.
72 . On these flagellants , see the summary of P. de Felice (1947, 204 - 7 ) and the long re-
views by Robert Hertz (1928 , 229 - 49 ) of K. K. Grass's book on Russian sects.
73 . Cf. M. de Certeau (197 0 ), the chapter "Le theatre des possedi:es" and p. 2 89.
74 . Although -si certainly stems from -asi meaning "female ," "wif e," in vod unsi , as in var-
ious other words formed in the same way, the suffix -si acruall) ' indicates a sta te of depen -
dence rather than the sta te of being a wife. Vodu nsi , in fact, designates boys as mu ch as girls .
Howe,·er, this is not the place to discuss this point in detail.
75 . "Individual" in the sense A. Zempleni uses the wo rd (1966, 316 ), meaning "one
whose starting point, origin al cause, lies in an individual problem," but nevertheless affects
the "familial unit. "
76. Hersko vits (1 938, 2:146 - 47 ) gives an example obse rved among the Fon. In th is in-
stance the child -to -be was consec rated to Sakpata.
11. 77 . Nago: a subgroup of th • Yoruba.

n J
78. For man y reasons I think that among the Gun (as among the Fon ) this recruitment as
a consequence of a "history of illness, " as Jane Belo would say in the case of Bali, or, by
personal vocation, one might say, in oppositio n to "impersonal vocation," is to be found
alongside the other. Bue it seems certain th at it by far the less frequent. Akindele and Aguessy
do not even mention it in the passage the y devote to this aspec t of vodun wo rship (1953 ,
113 - 14 ).

;:.!li:.J.i __ _
Notes co Pages 38- 45 331
79. In Africa at least , though not in Brazil. Cf. G. Cossard (1970 ).
80. Michel Leiris (personal communication ).
81. On this point , see G. Rouget 1961, 5 n. 9. According to Abbe G. Kiti (1968, 28 ),
three dialects "very different from one another" are learned, each being used exclusively du r-
ing various periods of the initiation. "Speaking in tongues" in Christian trances ob viously is a
practice to be linked with the use of secret languages. On glossolali .1, see the article on the
"don des langues" in the Di ctionn aire de sp iritu alite (1976 , vo:. 9 , cols. 223-27 ).
82. Concerning the use of the word "crisis" (for crise) in the present context , the siruation
is slightl y d ifferent in French from what it is in English, in which the existence of the word
"fit" sometimes makes things a lirtle clearer . A few examples chosen at random will suffice to
show that , as a whole, however , the problem remains quite comp arable in both languages . In
the introduction to her book, Jane Belo, who often uses the word "fit " (196 0 , notably pp.
223 - 24 ), writes (ibid , 4 ) about Balinese "occasional trancers " that "it w as sometimes possi -
ble to obser ve in these crises [italics are mine ] of involvement an analog y to sexual excite-
ment ," while Rob in Horton, speaking of the Kalabari in Afric a , refers (1969, 24 ) to "times of
communal crisis" when the priest was "apt to become spontaneously possessed, " a quite dif-
ferent use of the term, as we can see. On the other hand , "crisi s," which is an important entr y
in the index - from which "fit" is absent - of Religious Mov ement s in Co ntemp o rary Amer-
ica (Zarzetsk y and Leone, eds. , 1974 ), refers to trance in general but also to "crisis cults ,"
which perfectl y illustrates our point.
83. Loa bossal, by analogy with bossal, "slave," a term that during the da ys of the slave
trade in H aiti designated a slave recentl y imported and not yet baptized (M. Leiris, personal
communic ation ). The word itself derives from the Spanish basal meaning "wild, untamed"
(Metr aux 1958 , 326 ).
84. A. Zempleni (personal communication ).
85. Cf . S. Tidj arij ( 195 0, 299 - 305 ) and P. Verger (1 957, 95 ).
86. In the zar cult "the agitation of the g urri " is comparable to the mo vements of a slaugh -
tered chicken or an animal fatally wounded by a hunter as it struggles before its "soul comes
oucofit " (Leirisl958,81,n. l ).
87. Not to be co nfused with b or i, the possession of the Hausa.
88. 5inri: genii of illness who possess the sick person.
89. In other words, that of the priests themselves .
90 . Bia, a notion that includes simultaneously "shadow ," "reflection," and "soul" (Rouch
1973 , 531 ).
91. Later on , we shall see what, according to the myth , caused the first crisis incurred by
the first totemic priest.
92. Thi s will be dealt with later, in the chapter on exorcism.
9 3. This is curious ly similar to the ritual for Sak pa ta in Dah omey. After the adept has
been pub licly resurrected, the worms that have fallen from the shroud enveloping the
"corp se" are carefull y collected (Tid jani 195 0 , 301 ).
94 . M aximilien Quenum (1938, 93 ) gives another interpretation of this episode. Accord -
ing to him, the novice s "regarded themseh·es as having been forcibly torn from the delightful
co mp any o f their gods, " and this is wh y the y go into "a furor " and smash everything in their
path. Acco rding to Segurola (see under ah uan -ul i), who is clearl y using Quenum's text, this
ep isode form s part of the ceremonies m arking "the end of a sta y in a fetishist convent." See
also Le Heris se 1911, 135.
95. On this point, see R . Bastide 1972 , 94.
96. For the Wolof , cf. the film Le N 'doep; for the Mundang, cf. an unreleased film by Al-
fred Adler , Marie -Jose Pineau, and Andras Zempleni.
97. Cf. M. de Cerreau 1970, 61 and 207.
98. Lime, Dict io nnaire de la langu ef ranqaise ( I 978 ), s.v. "obsession:'
 ~!l
I;. 33 2 NOTES TO PAGES 45 - 53
1
)·1·
j,1
1P1l
~ 1 i
99 . Cf . Segurola 1963, unde r vodun. T he autho r adds the follow ing comment to his
transla tion : "T erm employed to convey that the person in ques tion believes himself to ha ve
I
received a wa rn ing from the fetish that he has been pe rso nally selected to enter the feti sh's
service
100. Still unp ublished , unhappily; let us ho pe tha t it will be soo n.
101 . In Portu guese chamar o ere (Cassa rd 1970, 164 ).
102. Metal bell, a sac red ins tru ment .
103 . On the ver y d isp uted origin of the wo rd ere, see Bastide 195 8, 199 n . 72 and 19 72 ,
62 n . 3 .
104 . 1 am quoting on the basis ofBasti de (1 958, 184 ).
105 . "D uring init iation the person in the santo state is mute and expre sses him / her self by
signs only"; Gisele Cassa rd (personal com municat ion ).
106 . Readi ng the text, one gets the impr essio n tha t the novices return ing from their bath
are singing in the santo sta te, whi ch would cont r ad ict the info rma tion given in th e pre vious
note, that novices in the santo state are mute . Clearly there is a sligh t erro r in th e wo rdin g of
the description .
107 . Omotun may well mean "child once more," implying tha t the novic e has retu rn ed to
the child state . Ab raham's Dicti onary of Modern Yoruba ( 1958 ) indicates that tun exp resses
the idea of beginn ing aga in rathe r than "newness," but this is a point that needs fu rt her
checking .
108 . Cf. Bastide ( 1958, 199ff ) and P. Yerger (1957, 72 - 73 ).
109 . The god of lighting, one of rhe principal orisha.
110 . In Haitian voodoo, a direct descendant of vodun worship, initiation is much briefer
(cf. Metraux 1958, 171 - 88 ) and does not occu r at the same stage of the ade pt's life. Simpli -
fving a great deal, one might sav that it is more a confirmation than an initiat ion . In H aiti,
the adepts are not trained to the practice of possession from ea rly childhood on, cont r ary to
what often happens in Benin. T his in part explains why initiation is so different on rhe rwo
sides of the Atlantic . Even so, Haitian initiation does present a numbe r of featu res th at con -
form to the African model.
111. On initiation into the uodun cult, see Herskovits (1938, 2 :lllff. ), Akindele and
Aguess, ( 1953, 113ff. ), Yerger (1954, 322ff .), and Kiri (1 968, 26ff. ). Nowadays an agree -
ment "·ith the Marxist -Lenin ist government of the Popular Republic of Benin has fixed the
period of initiation at three months at most.
112. Although this word, now used tradirionallr in this context in French, can give rise to
confusion, I shall continue to use ir here to designate what is called in Benin, in Fon - and
Gun -speaking territory, hunkp ame (the god's enclosure ) or hunxwe (the god's house ) or
ki1xwe (the house of death ), in which hun (blood, divinirY, secret ) is in practice synonymous
with vodun, and in which ku denotes the symbolic death that marks the beginn ing of initi-
ation.
113. Or , more fully, rhe "new wi,·es of the di,·inity,'' the huns i yoy o, bur also "husbands "
since in this context -si denotes dependence without sexual connotation, and applies to boys
as "·ell as girls.
114 . I refer to them as being female since girls were in the majority, but there we re a few
boys as well.
115. I say "presumably " because I can't rule out the possibiliry that these "new wives" m ay
have gone elsewhere at a later date in orde r to perform the same ritual. T he fact rem ains that
~,·
h
I
!·
once their initiation is ove r, the new initiates will neve r sing this song agai n.
116. About Siligbo, see Akindele and Aguess y 1953, 123 .
:~ 11 7. With one difference : Le Herisse is desc ribing three new adepts of w hom only one is
in the ,·ery first stage of initiation, while the othe r two are at a slightly more adva nced stage,
whereas Verger's photograph is of four adepts all at the very first stage .
118 . Gisele Cassard 1970, 172 - 74, 181, etc.

I·
ii -~L _
Notes tO Pages 54- 66 333
119 . T he film Sortie des nouvelles de Sakpata (Ro uget an d Ro uch 1963 ) includes in its
second part the first of these ceremonies, which consists in going to tak e a ritu al bath in th e
sacred stream . Jean -Luc Magneron's film Vaudou shows seve ral aspec ts of initi ati o n in Da-
homean convents, but in a context so confused and falsified that this piece of evide nce , ho w -
ever beautiful to look at on occasion, unfortunately remains of little use.
120. Cf. Rouget 1961, 3- 4. The fir st part of the film Sortie des nouvelles de Sakpata
shows the novices performing these begging dances, followed by givings of thanks, but be -
fore their official graduation and within the convent enclosure.
121. Cf. the description of it by Herskovits ( 1938 2: 165 - 66 ). This ritual was filmed for
Fr ench tele vision in 1971 by J. Lallier and M. Tosello, in col la bor at ion with P. Verger . It is
the same rite that is observed, with variations, in the candomble as described by G . Cossard
(1970 , 189 a nd 143 ) and in H ai tian voodoo by Merraux ( 1958 , 181 - 87 ).
122. Which of course does not form part of the category of"song " in the Fon and the Gun
thought system. But I shall leave this aspect of things aside, important though it is.
123. Thi s repert oir e is the subject of a study, in preparation, by the present author.
124. More precisely for the vodun Sakpata of a particular villag e .
125. Do '(la ' = 440Hz ).
126. For reasons that would take too long to cover here, it has so far pro ved impossible to
issue a record of these "long thanksgiving songs." Howe,·er, some examples of the sho rt rype
have been issued (Rouget 1961 ); it at least gives some idea of the strangeness described here .
12 7 . I do nor wish to conceal the fact that there is an element of interpretation in this way
of presenting thing s. Once again, they are not well known. T he interpret at ion offe red here is
based on the hypothesis that what happens in vodun worship is compa rable to what has been
observed in orisha worship, and particularly in the prac,ice of the candomble. It has the mer -
it of providing an account of a number of facts that would otherwise remain incomprehensi -
ble. It nevertheless still requires verification.

CHAPTER T \X"O
l. Thi s is a point about which a great deal has been written during the last ten yea rs. It
seems, however, that the last word is very far from ha ving been said. On this subject, see in
particularK. P. Wachsmann ( 197l ) andB . 1 ettl (1977 ).
2. Recent rese a rch has re vealed the scope and complexity that such a taxonomy can attain
in a Melanesian society (cf. Zemp 1978 and 1979 ). At the moment, however , research of this
kind is still exceptional.
3 . See note below.
4. In line wi th J. Molino ( 1975, 47 ), J. - J. Narriez, in his work on musical semiology
(197 5 , 50ff .) proposes that musical analysis should be divided into levels - po"ietic, neutral,
esthesic - likewise inspired b y linguistics. For various reasons (the choice of terms in the case
of "po"ietic" and "esrhesic," the validiry of the concept in the case of the "neutral " level ) this
terminology does not seem to me to offer sufficient advantages (although it does have some,
incontestabh· ) to be preferred here to that proposed by J akobso n.
5. This is the case "·ith the tr ance observed by M. Leiris ( 1934 , 114 ) during a hunter's fu -
ner a l rite. '"The aspira nt guardian of rhe family totem" is described as turning somersaults,
rolling in the dust, and standing o n his head, after which he is "led away in convulsions."
6 . Germaine Dieterlen and Amadigne Dolo (Dog o n of Sanga ); personal communication .
A phorograph published by G. Dieter len (1941, pl. XIV ) shows a future binu priest hanging
b, · his feet from the roof of a ba rn , which he has climbed during his trance . Apart from the
binu priests, both priests a nd priestesses of Amma (god of cre ation ) on the one hand (cf. Die -
terlen 1941, 240 - 45 ) and certain "wives of the Sigi" on the other (cf . Griaule 1938, 268ff .;
Leiris 1948, 13 lff., 353ff. ) a re also subject to tr ance . Onl y the trance of the binu pi rests (cf .
Dieterlen 1941 and de Ganay 1942 ) is described as regularl y associated, after "ordination ,"
with a set musical repertoire. I am tempted to think, for va rious reasons too complicated to
334 NOTES TO PAGES 66- 75

go into here, that the trance of the binu pr iests alone ought to be seen as falling int o the cat e-
gory of possession in the sense it is used in this book.
7. Cf. p. 98.
8 . Griot : a West African French word meaning a professional musician belonging to acer -
tain caste.
9. "T he greatest of all the bori gods" (Monfouga - 1icolas 1972, 105 ).
10 . And r as Zempleni (pe rsonal communication ).
11. A female ndop initiate .
12 . If one had to compare bori, candomble, and vodun wo rship as systems develo pin g in
rime, then one could say that in the bori it is possession that counts prima rily, wit h ini tiation
playing only a limited role, that in vodun worship, on rhe contrary, it is initiatio n that cou nt s
the most, while possession occupies a statistically small amount of time, an d that the can-
dombl e rep resents a sort of compromise between rhe rwo, since initiation in the candomble is
more important than in the bori, and possession is more frequent than in vodun worship.
13. Cf . pp. 43 and 150 - 51 .
14. Cf. p. 47.
15. T he woman in charge of rhe initiation of no vices alongside the "fa ther of the saint ."
T he latte r is the cult's principal priest and head of the religious com munity. The adep ts are
the "sai nt 's daughte rs." In this syncretism characteristic of the candomble, the wo rd "sa int"
(Portuguese santo ) designates the divinity, the orisha.
16. A double or triple bell with internal clapper that only "the fathe r of the saint, or per-
sons carrying out particular duties, have the right to shake" (Cassa rd 1970, 70 ).
17. With reference to "wild tr anc es ," R. Bastide (1958, 252 and 1972, 94 ) observes char
in the candomble there are "leathe r chants ," a term alluding to the leather -thonged whips
whose ultimate aim is "to create a violent ecstasy in those guilty of some taboo violation or of
some lack of respect. " In practice, the "savag ery of the crisis " trigge red by these chants has no
connection with the "wild" possession we are dealing with here. So the example of these
"leathe r chants" does not run counter to the observations I have just made.
18. P. Verger (personal communica tion ).
19. This can be both seen and heard perfectl y in the film shot by P. Verger and Y. Bellon
for French television in the Sakere region of rhe Popular Republic of Benin in 1974 .
20. Analogous cases of conventions have been reported in Bastide (1972 , 93- 4 ) for the
candomble in Brazi l, and in Blacking (1973, 44 ) for possession among the South African
Venda.
21. Adler and Zempleni (1972, phot o facing p . 129 and personal communication ).
22. Jane Belo's descriptions are always so painstakingly derailed that one is jus tified in
chinking that if there had been instruments involved she would have said so. T he fact that the
music is purel y vocal is corrob orated by recordings of sanghyand dedari and djaran (nymph
spirit and horse spi r it) in which one hears only choral singing , mingled with thos e rhythmic
cries made famous by the ketjak (cf. G. Fresna is, Bali, discography ).
23. Cf. p. 108.
24 . I shall deal with the dhikr at length in part tw o.
25. In chapter 5, " Music and T rance among the Greeks ."
26. Pizziche terantate, literally "tarentian stings," is the term that metaphorically desig -
nates the rune to which the pers on bitten by the tarantula must dance, so closely associated
are the music for the spider and the one for its bite.
27. Bells: nor our church bells, of course. I am thinking here above all of the simp le or
double iron bell, held in one hand and struck with a wooden stick held in the other, which is
so common throughout black Africa.
28. Fiddle: any instrument with bo wed string (s) and a neck .
29. I shall return at length to the drum in a later chapter, pp. 169- 76.
Notes to Pages 76- 98 335
30. Cf. lacer, in cha peer 5, "M usic and Tr ance among the Gree ks."
31. Samuel Baud-Bovy has writte n co me, for which I am very gra ceful, "chat the instru-
ment Dionysus plays is nor the Apollonian lyre bur che instrument, of a quite different type,
which Wegner calls barbicon. This barbicon occu rs only in depictions of banque ts , as an ac-
companiment co drinking songs, and in the hands of Eastern Greek poets, Alcaeus, Sappho,
and Anacreon . ... " T he lyre mentioned in che next example, howev er , is, as I have indicat -
ed, the classic five-stringed type .
32. Here I am talking about the "governm ent oracle," which is the principal Tibetan ora-
cle. For oracles of lesser importance it seems chat these horns are not blown . In these cases a
bell or cymbals are used (Nebesky -Wojkowirz 1965, 435 and 438 ), sometimes acco mpanied,
as rhe horns are, by drumming.
33. Cf. p. 155.
34 . In the cult observed by P. Simon and I. Simon -Barouch (1975 , 40 ) there was no lure
because there was no one there co pla y ic and the drum was not of the required rype.
35. Cf . Moursopoulos (1959, 131 n. l ) andjea nm aire (1951,37 ).
36. As transcribed (p. 181 ), the arebate rhythm includes bursts, if one may so put it, of
demisemiquavers at crotchet= MMl 12 . In ocher wo rds, the smallest beat unit of this rh yth -
mic theme is MM 896, or very close to fifteen strokes a second, which does seem extraordi-
narily fast, even wh en there are, as in the case here, intermittent and very brief bursts. Ir is
doubtful chat this kind of rapid cadence can be done with bare hands; arebate must, there -
fore, be drummed with sticks.
37 . On the adarum rh ythm, which certainly deserves a study of its own, see P. Verger
( 1957 , 22 and 243 ). T he adarum can be heard on the candomble record made by G. Behague
(see discography ).
38. My italics.
39. My italics.
40. Ac lease this is what emerges from the recording by L. Berthe, available in the Musee
de !'H omm e collection. See discography.
41 . A double iron bell with external clapper, pare of the orchestra and different from the
sacred adza mentioned ear lier.
42. Unpublished films and recordings .
43. One of them published; see record Ogoun, dieu de fer.
44. Seep. 106.
45 . By "rh ythmics" I mean the ensemble of the rhythms and the rhythmic traits characte r-
istic of an ethnic group or a region, as well as of a musical sryle or a composer , the rh yt hmic
system being char whi ch rules the rhythmics.
46 . Nago: one of che Yoruba dialects.
47. John Levy, notes for the record Musique bouddhique de Coree (see discography ).
T his is the ch ant called "T he Peak of the Vultu res ," a sore of requiem intended "co prepare the
soul of the departed for its entry into paradise ."
48. This is also made plain by the music of the film about chis cult made by R. Morillere
and the two writers (see filmography ).
49. On possession among rhe Zu lu, see S. G. Lee (1969 ).
50. Ndopka t; female initiates co the ndop.
51. Cf. p. 66.
52. Binu boy "drum of the binu," nor co be confused with binu boy (high cone), meaning
"name of the binu." See G . Calame -Griaule 1968, 36.
53. Diecerlen 1941, 224 n. 1, and 226 n. 4 .
54. Nommo: one of the two primordial twins created by Amma, rhe god of creation.
Nommo is the master of water and words. He is the focal center of the binu cult. See G. Ca -
lame-Griaule 1968 , 203 .
 336 NOTES TO PAGES 99- 11

)), On the importance of the musical or verbal "motto" and its "powe r of evocation" !
black Africa among the Peul, outside the possession context, see Christiane Seydou (197:
28 ).
56 . Unless these dance rhythms are functioning as collective mottoes for the gods' T his
merel y a hypothesis; one cannot tell with so lirrle data .
57. On the Sardinian argia, see Clara Gallini's book (1967 ).
58 . Th onga and Tsonga are two different transliterations of the same name. Tonga , ho\
ever, denotes a quire different ethnic group.
59. The mottoes are frequentl y played on an instrument by a musician and / or sung l
the musicants.
60. Let us leave aside the spectators, whose participation in the music is secondary ,
least from the point of view of what concerns us here.
61. Thi s ceremony has been described in detail by P. Verger (1957, 150 - 58 an d phor ,
gr aphs 83- 86 ); che music has been published in disc form by G. Rouget (see discography ).
62. This ceremony has also been described in derail by P. Verger (1957, 316 - 21 ) an ,
from another angle, by G. Rougec (1965, 70- 76 ); the music has been issued on disc (s,
discography ).
63. Marion Kilson does not specify at what moment in the ceremony these songs a
sung, or which of chem might be used in chis context to summon the gods and trigger posse
t sion. The collection of songs she published (a considerable number: 243 ) is extremely ht
i: erogeneous; four -fifths were not collected in their siruacional context, and only three were
'·
; "i face sung by adepts in a state of possession. le is therefore difficult to say what relation che
may be between these songs and possessi on as such.
64. \'iiadiiga: possession seance.
65. The dhikr will be dealt with at length in part two of chis book.
66. Here , the lease one can say, however , is that the music triggering the possess ion rran
is nor ex actly made up of the genius mottoes! Moreover, the behavior of the possessed is ve
particular: "\'(!hen , ange red by che repecinon of God's name," Sophie Ferchiou writes, "ti
genius responsible for the possession wishes to manifest its anger, one sees the 'patient ' be
her thighs or scratch her cheeks like a mourning widow."
67. This air will henceforward be his, H. A. Junod notes (1936, 2:441 ), and adds ch
chis song will lacer be the means by which his crises are pro vok ed or cured. He does not ma :
clear, however, whether che song is sung by the possessed person himself in order to indu
or heal these crises .
68. A. Zempleni (personal communication ).
69 . See pp . 24 8- 49 and G. Rougec 1955 (discography ).
70. In the context of divination, not possession, Germaine Dieterlen (1951, 221 - 22 ) re
ords an identical fact among the Bambara, where it is the diviner's ha rp th at is the commun
eating instrument.
71. The information was acquired ,·ia a \Volof interpreter, and Michel Leiris himself fee
it should be created with caution . With the pro viso chat it requires verification , however,
certainly merits inclusion here in view of its interest in relation both co the water drum 's syn
bolism and its organology.
72. The iron bar is probabl y arranged slantwise in the water container; the lip of the UJ
rurned gourd, which floats on the water and is struck, rests aga inst the bar, which is remini
cent of the positioning of the gourd and rod in the gourd drum of the Songh ay (Rouch 141
Surugue 1972, 52 ), which is also a possession drum but rests on sand rather than water. /
among the Malinke (see below, p. 116 ), who also use the water drum for possession dance
among the Bambara , the instrument is beaten by women.
Among the Koniyake, a subgroup of the Malinke, the water drum is beaten by men ar
used for profane dances performed by girls (Rougec , discog raphy 1972 ). As we know, in ti
Popular Republic of Benin, Cuba , and Brazil, it is an instrument used for funerals, and it
Notes ro Pages 114 - 26 337
played by men. We thus see to what a gre'a r extent the symbolism of an instrument as specific
as the water drum can vary from one region to anothe r : an example - if one we re needed - of
the largely arbitrary nature of this kind of representation.
73 . I have done no rese arch into what ma y have been writt en on rhe relation between
dance and possession . Let me simply observe here rhar in her "Panorama of Danc e Ethn o-
logy," which is by no means new since it dares from 1960, Gertrude Kurarh gives in all (p .
240, Psychology ) only two references concerning possession: Sachs and H erskovi ts.
74. M. Leiris gives a much more precise indi ca tion of what the gurri is in a lat er publica -
tion (1974, 43 ): rhe gurri (from rhe verb agworra , "to low, to roar, to weep whil e howling" ) is
a form of behavio r typifying rhe possessed person and a conventional sign indicating the ar ri-
val, then the depa rture of the spi rit. Ir consists in violent movements - rotatio n or swi ng ing
of rhe upper torso - accompanied by noisy exhalation and rhythmic gesticul at ion , which the
majo rity of zar are supposed to have those who m th ey put into tr ance execute, and whic h dis -
plays different for ms depending on the different categories of zar.
7 5. "A man divided, rhar may well be the origin of the theat ric al character" A . Schaeffne r
writes ( 1965, 42 ) in his study of prerh earer.
76 . A vast subje ct rhar could only be adequa tely trea ted , of course, in an interdisci plina ry
manner. From the overall standpoint taken in this book and to speak only of Fren ch sources,
no sati sfactory wo rk has been devoted to it so far. T he book by Robert Frances on rhe per -
ception of music has quire different concerns . Th e author himself pre sents ir in fact (2nd ed .
1972, 414 ) as co nstitutin g a "proleg ome na ro a the ory of esrheric judgment" and as using the
data of experimental and social ps ychology (ibid, 402 ), and ignoring , or almost, those of
physiology, while making no use of comparison.
77. Here I am following A . Caraperyan 1948 , 148.
78 . A very ancient form of divination practiced by the T ibetans consists in reading the dis -
placement of seeds placed on a drumskin when the latte r is caused to vib rate under rhe effect
of a nearby drum beaten by the soothsayer, who simultaneously invokes the app rop riate di-
,·iniries and requires them to answe r his questions ( ebesk y-Wojkowi rz 1956, 457 - 60 ).
79 . On thi s topography of ou r inne r sensiriviry, see H usson, La voi x chantee, p. 60.
80. A particular theory abou t the neurophysiolog ical effects of drumming will be ex am-
ined in chapter 4 , "T he Strange ]\!echanism. "

CHAPTER THRE E
1. Although they lie outside rhe scope of this book since they deal with New Guinea , I
would refer rhe reader here to rhe six articles dealing wirh posses sed per sons, mediums, and
shamans edited and introduced by B. Juillerar for rhe Journal de la Societe des Ocea11istes,
1977.
2. Seepp.102 -14 .
3 . Cf. in M . Eliade's work (1968 ) the principal references to music on pages 111 - 12,
115 - 16, 144 - 53, 161 - 65, 234,237, 243 - 45, 25 7, 390 - 91. For rhe Tungu s in particular,
see rhe analytic bibliography in L. Delaby 1976.
4. On rhe music of the Araucan shamans, see Maria Ester Grebe 197 3 .
5. Information kind Iv supplied by Laurence Dela by.
6 . On rhis point, see E. Lor -Falck 1961.
7. Rob erte Ha ma yon, accordi ng to Zangalov (personal communication ).
8. 1ores to rhe record Seance de cha manisme. Recorded by H. Le Besnerais . Texr by S.
Dreyfus (19 54 , 2). See discogra ph y.
9 . L. Dela by and R. Ham a yon, ro whom we owe the posthumous publication of rhe wo r k
of Lor -Falck, refer ro Vasilevic ( 196 8). T he origin of rhe wo rd shaman has, of course, been
rhe subject of a great many different hypotheses.
10 . T he third means "inhaling the auxiliary spirits, curing rhe sick, searching for the pa-
tient 's spirir," and is therefore of less inreresr for us here than rhe first rwo.

........
.iJ.
 11
·!
I
:ll::f !
338 NO TES TO PAGES 126- 31

I: 11. I am thinking here of the Sanskrit wo rd san:zgita , which includes vocal music, insrru-
, I
mental music, and dance (Bake 1957, 196 ).
12. A derail kindly supplied by Lau rence Delaby (personal communication).
13. The bibliography on rhe shamanic drum is considerable. See M. Eliade (1968, 128,
144 - 45, and 390 nn. ).
14 . Italicized in original. Eliade is here contrasting musical magic and noise magic, with
reference to the use sometimes made of the drum to "drive out the evil spi rits ." We shall re-
turn to the subject when dealing with exorcism.
15. "The drum speaks," the Tungus say (Dela by 1976, 109 ).
16 . We will deal at length with this problem, which is very important, at the end of the
first part of this book.
17. See note 30 below .
18. See Anne Chapman's notes for the two records of Selk'nam sha manic songs she has
i: published (discography ).
19. Shirokogoroff constantly uses the word ecstasy, never the word trance. What he is de-
scribing is ne~•ertheless what I proposed to call "trance" in this book.
20. Cf. in particular ibid., pp. 326 and 309, the first passage concerning a Tungus sha-
man, the second a Manchu shaman.
21. Ibid., p. 006 .
f;
22 . I imagine Shirokogoroff means alcohol here. Elsewhe re (1935 , 306 ) he describes the
t'··~ shaman as swallowing a large shot of vodka before setting our on his descent to the lower
~. world .
23. In consequence, it is never played outside the context of a shamanic ceremony, since it
could provoke the inopporrune visit from spirits "who would then seize upon people incapa -
ble of mastering them" (ibid ., 302 ).
24. Information provided by Anne Chapman, who acqui red it both from her own person-
al research in Ti erra de! Fuego and from reading Gusinde.
25 . 1 • ores ro the rec ord ing made by Le Besnerais and based on evidence collected by him .
See discography.
l. 26. The song can rake on yet other aspects . Among the Buryars, for instance, unlike the
shaman's invocation song which "obeys musical rules," the "song for the journey is a form-
less and continuous sequence of sounds, intersected by moans, sighs and cries (copied from
birds, creatures that can pass without hindrance from one world to another )." See R. Ha-
mayon 1978b, 24.
27. See discography. One side of the rec ord is devoted to the "Hol y Ascension," in other
words rhe shaman 's journey to the upper wo rld . Unforrunately, however, this is nor a field -
work document bur a recons tructi on made for the purposes of a "folk music concert" in the
U .S.S. R. in 1969. T he violent and frenzied side of the trance so often described in relation to
Siberian shamans is natur ally absent. This said, the remainder is certainl y very representa-
tive of Yakut shamanic music. Being sevenry -eighr yea rs old, but of incredible youthfulness,
the singer, S. Z verev, clearly remembers the rime when shamanizing was still common. One
presumes he sha manized himself. Even down to the shaman 's great yaw ns as he incorporates
the auxiliary spirits, one might think that this side of the reco rd was specially made to illus-
trate the two shamanic journey narratives reproduced by Eliade (1968, 189ff. and 160ff. ),
one from rhe Yakut, one from rhe Altai Tart ars.
28. Side B, band 9 of J. Huraulr's record. See discogra ph y.
29. Ir has nor been possible to establish rhe text of most of the othe r eighteen , which
makes the proportions of the first sixteen even more significant.
30 . On the sryle of these songs and its kinship with that of songs among Siberian tribes
(Tungu s, Chukchi, Yukaghir, Samoyed, Osriak ) mentioned in this chapter, see A. Lomax
and his "cantomer ric" analysis at the end of Anne Chapman's notes , as well as his book Folk
Song Style and Culture. Ir is undeniable that the evidence he invokes must be taken into ac-
Note s ro Pages 132 - 48
339
count . It fits in perfectly wi th Levi-Srr auss's observa tions (1966, 326 ) on "th e an cient pal
lithic cultu ral srock" co m mo n t0 Asia and the New World. eo-
31. Cf. p . 17ff.
32. T his is the case in particular w ith the descr iption Avicenna made of a shamanic divi-
narory seance among the T urkoman . Cf. J .-P. Roux (1963, 290 ), in wh ich cert ain quotatio ns
from ancient texts, used in his article on shamanic dance, are very insrru ctive .
33. Seep . 28 .
34. See 1 Samuel 10 :5- 6. We shall return t0 this late r.
35 . See 2 Kings 3 :10 - 15 .
36. Le theatre sacre des Cevennes cites innumerable examples (M isso n 1707).
37. On divinarory tr ance in shama nism, see E . Lot -Galck 1968 .
38. A medium presented by adel, wrongly in my op ini on , as a shaman.
39. Cf. p. 000.
40 . This subject will be discussed at length in pa rt rwo of this book .
41. More precisely, among the !Kung Bushmen, a gro up referr ed tO in the two studie s
quoted here. Let us recall rh_arthe Bushm en are peopl e of small statu re who live by hunting
and gathering in the Kalaha ri desert in sout hern Afri ca.
42. "Medicine man," "medicine song," and "me dicine dan ce" are the terms used by Lorna
Marshall. In reality, according tO Ma rshall's data, it wo uld be mo re proper t0 translate the
Bushman terms, more lite rally, as "power man," "powe r song" and "powe r dance" in which
"power" or "potency" would be a rendering of the wo rd n/ um.
43. Even when they do have this powe r, which is ext remely rare, the wo men do not act as
"medicine women" (Marshall 1969, 366 n . 1).
44. Although, as we have seen, these dances are no r always specifically aimed at he aling a
sick person - very far from it- th ey do neverthel ess, in all cases, aim at defen ding and pro-
tecting the individ uals and the group .
45. Fou r examples of medicine or healing songs, pr esenting diffe rent aspects of this mu-
sic, have been issued on disc with transcriptions (cf . discog r aphy ). See also the wo rk of 1-
cholas M. England (1967 ).
46. Cf. Bushman Music (see discog raphy ).
47. Cf. Seance de chamanisme (see discography ).
48. Cf. M usique . .. Wayana de Guyane (see discography ).
49. Recordings archived at the British Institute of Reco rded Sound , Londo n, and the De-
parremenr d'erhnomusicologie du Musee de !'Ho mme. Audrey Butt he rself says of these
chants ( 1962, 41 ) that the shamans explicitl y claim that the chant and its rhythm aid th e spir-
it t0 rake flight.
50. Bori, in Hausa, means "boiling ." Among the ancient Greeks, the Pyrhia was "rega rd-
ed as a living cauldron who boiled under the god's influence" (Jeanmaire 1951, 452 ).
51. Cult in the sense the term is used in this book, since, as is evident in Michel de Cer-
reau's work, the possession of Sister Jeanne des Anges and '1er companions was in fact int e-
grated into the mass as part of rhe pe rformance, and indeed became its "high spot ." Bur it fig-
ured rather like an anticulr. Moreover, Michelet, in his La sorciere, has shown the great
importance that must b~ given t0 sorcery in these affairs, somethi ng that removes them even
further from possession as l have defined ir.
52. lam speaking here of a traditional Dahomey, one nor rorn apart by conflicts between
the traditional religion, Christianity, and Islam. This Dahomey no longer exists. But its tra-
ditional religion is still sufficiently alive for it t0 be poss ible to ignore this aspec t of things
here.
53. One cannot quire rule our, however , the possibility that in vodun ceremo nies such as
those held for the graduation of the new initia tes- ceremo nies entirely cente red aroun d pos -
session - the role of the nonvito o r "she who calms the mothe r" is in fact to exo rcise malevo -
lent powers . We can ignore this point here, however, first because it is an enti rely personal
 340 NOTES TO PAGES 148- 67

interpretarion that remains to be proved, second because the intervent ion of rhe nonvito, in
these circumstances ar least (the only ones I have been able to observe, moreover), is so dis-
creet as to pass almost unperceived_
54. Is he in trance at this moment? What exactly does he drive our of the dead person's
house' Does the rirual require music? I am unforrunarely unable ro answ er these quesrions.
55 . The same siruarion can be observed in southern India. Cf. L. Dumont (1957, 353ff. ).
56. Junod's italics.
57 . As presented by R. F. Gray (1969, 171 ), the shetani cult among rhe Segeju of Tanza -
nia would seem to be an exception, since he describes it as aimed at "exorcis m" of the spirit
respons ible for the possession. In fact, everything indicates that here again the re is, at least,
pact making involved_ The use of the wo rd "exorcism" is thus no more justified in the case of
the Segeju than in that of the T honga .
On the other hand, the case cited by John Beattie (1969, 164- 65 ) among the Banyoro is
quire clearly a real exception and would thus seem to invalidate what I have said. In fact,
however, rhe proces s o_fexorcism, or rather of destrucrion (Beanie's term ) concerns a cate -
gory -of ;;pirits (ghosts ) rhar is very marginal in the ove rall picrure he gives of possession
among rhe Banyoro; and this overall picrure - including what he has to say about rhe use of
music (ibid, 166 )- is in perfect accord with what I stared here.
58. According to Sachs (1940, 106 - 8), th e instrument David played - the kin nor - was
neither a harp nor a zither, contrary to what is found in the many translations of the Bible,
but in fact a lyre.
59 . The wo rd "harp" here translates the Hebrew wo rd nevel. Sachs ( 1940, 115 - 17 ) con -
siders rhar it was probably a vertical angular harp. On the other hand, the instruments here
called "flutes" are in fact clarinets.
60. Dhorme observes in a note (19 56 , 1143 ) th at "the music is at the service of inspira·
I- cion."
61 . A fact pointed our to me by Elen a Cassin .
62. Particularl y E. De Manino. See his chapter "Ethnological and Folklore Pa rallels ."
63. The argia, its Sardinian equi vale nt, is now also well known, thanks to the work of
Clara Gallini ( 1967 ). Unhappil y, I became aware of it to o late to make use of it in this book
to any large extent.
64 . Our quotes refer to the French translation, La Terre du remords, published in 1966.
Let us recall th at the original Italian edition has an accompanying record made by Diego Car -
pirella who contributed the chapter "T he Choregrapho -Musical Exorcism of Tarantism,"
and made a film of a rarantism seance which was shown in Paris, in 1960, on the occasion of
the Sixth Internation al Congress of Anthropological and Ethnological Sciences.
65 . Elena Cassin (see bibliography ), in her account of this work, has shown the merits of
rhe book and rhe inadequacies of the main thesis.
66 . These data are taken from "Ne urops ychi atric Considerations on Tarantism ," written
by Dr. Giovanni JerYis and fearured in De Martino 's book as an appendix.
67. Muzak, no. ':58 - 59.
68. The Gnawa will be discussed in chapter 7.

CH.~PTER FO UR
l. In various passages in the article "Music," particularly rhar (1768, 314 ) in which, hav -
ing cast doubt on rhe fidelity of transc ripti ons, he poses with great lucidiry rhe problem, cen -
tral to erhnomusicology, of the inte rpretati on of musical systems.
2 . Th e passage of the Diction naire quoted here alreadY appears, under a nearly identical
form, in rhe article "Musique " of rhe great Encycloped ie of Diderot . Of his Dictionnaire,
completed in 1764 and published in 1768, Rol'Sseau writes, in the first page of his "Preface,"
that "rhe bases of this work were quickly set forth fifteen yea rs ago in the Encyclopedie." He
furthe r stares, on th e last page, that its manuscript {which was composed in three months )
Notes ro Pages 168 - 79 341
"left his hands" in 17 50. It is thus in 17 50 that the first draft of the above mention ed passage
of this Dictionnaire was wrirten.
For its part, the Essai sur l'origine des langues, ou ii est parle de la melodie etde /'im itation
musicale appea red in 1781, afte r the death of Rousseau . The date of its composition has not
yet been ascertained and there has been much dispute about it. According ro C. Porse t (1970,
15 ), the Essai was sketched out "during the writing of the second Discours [in 1754, conse-
quently ] . . and resumed in 1756, perhaps as an Essai sur le principe de la melodie ." In the
Confessions, Rousseau writes that in 1761 he was still working on the Dictionnaire, but tha t
he had just finished the Essai. So it is between 1750 and 1761 , at the latest, that Rousseau
changed his mind about the effects of music . We still need ro know when exactl y Rousseau
wrote the text concerning the division of music into two kinds, natural and imitative, a text
that conciliates his two theses , that of the physical and that of the moral effect of music, or
rather that shows how each completes the other. T his text appeared in the 1768 edition of
the Dictionnaire and then in 1782, in the Supplement a l'Encyclopidie .
3. My italics.
4. In the Dictionnaire, page 308, at the end of the passage concerning the distinction
made by Rousseau between natural and imitative music, one can read: "As long as we search
for moral effects in the ph ysics of sounds, we will not find them there and we will discuss
without hearing one another," a sentence which, obviously, prefigures the one of the Essais
cited abo ve.
5. Essai sur /'origine des langues ou ii est parl e de la milodie et de /'imitation musicale (ed.
Porset 1970, 179 ). Rousseau is clearly repl ying here ro the writings of a certain Burerte, who
held the opposi te view, then very fashionable. Porset reproduces a significant passage from
Burette 's wo rk in his edirion of the Essai, p. 164.
6 . Porset 1970, 165.
7. See the section "Wha t Music?" in chaprer 2.
8. To be sure, eedham's thesis is presented here in a much abbreviated form.
9 . Needham refer s here ro three articles on percussion instruments that appeared in the
En cyclo paedia of Religions and Ethics (1912 ), and particularl y that by Crawley ("Drums
and Cymbals" ), of which elsewhere ( 1967, 609 ) he praises the psychological and "even neu -
rologic al" orientation .
10. Primed rhe following year in Man, n.s. 3, no. 2 (1968 ): 313 - 14 .
11. :'leedham (1 967, 613 n . 1) is here using obs erv ation s made by Francis Hu xley. I shall
return later ro these perturbarions of the inner ear.
12. We shall return to the word "driving" in the next section.
13. Among those who have been con vinced by Neher 's thesis, I should also cite G . De ver -
eux (1 970 , 58 ), who cites it as aucho ritative in his general interpn ration of trance, and, more
recentl y, F. Schott -Billmann ( 1977, 140 ).
14. Quotation taken from Neher (1962, 155 ).
15. It should be noted that a germ of this theory of the condirioned reflex already existed
in a remark made by ina Rodrigues, who, parallel with his inrerprecarion of hypnosis, also
wrote: "There are initiates who cannot hear the music or song with which their first Santo
state coincided ... wi th our that scare manifesting itself" ( 1935, 111 ).
16. T aken from P. Verger 1957 , 76.
17 . Is the music powerless without the reinforcement of a special set of social circum-
stances' Blacking asks himself (1973, 45 ). For him, the answe r would be, on the contrary,
chat if music remains without effect in certain cases this is due ro certain social inhibitions .
18. M y italics.
19. A weight technique also used, in different forms, by Tibetan oracles and Siberian sha-
mans.
20. Misprinted as "psycholog ical" in the published text. The context indicates that it
should be "physiological," and the au th or has confirmed this.
 342 NO TES TO PAGES 1 79- 94

21. G. Cossard (personal communication ).

22 .. Abo: a potion made of various plants with "the property of pro voki ng and rein forcing
the tr ance state " (ibid, 181 ).
23. Zempleni 's thesi s consist s of more than this single consideration of the music's effect ,
however. He lays stress thro ughout his wo rk on the importance of the bak or sung mottoes.
24. Italicized by Dr. Aubin (194 8). Qu otation ta ken from Leiris (1958 , 18, n. 2).
25 . Cf. his various films on possession.
26. Cf. B. Surugue (1972, 49ff. ).
27. See pp. 128ff.

CHAPTER FI VE
-1. No referenc e to a particular edition of Plato 's wo rk will be made here since the quota -
tio ns given in thi s book come from different sources. Some were borrowed from the Loeb
collection or from translations br Jowen, H amilton and Cairns, or by Dodds. O thers were
translated d irectly from the French (Editi on des Belles Lem es: Collection Bude ) which is of-
ten closer to the Greek text. Still ot her s we re translated by Eric de Dampierre an d myself.
T he same applies to the oth er Greek autho rs cited.
2. Concerning the tr anslation of psyche by "soul," see Dodds (195 1, 138 - 39 ).
3. T he most ancient usage of the wo rd enthousias mos is known to us through Democ ri-
tus, who gave "for the first time a really scientific and rationalist explanatory essay of enthu-
siasm," writes Delatte (193 4, 33 and 28 ) in his stud y on the conceptions of enthusiasm
among the Presocratic philosophers .
4. Th at En theos should be translated in this way is stressed by Dodds (1951, 87 , n. 41 ).
He points out tha t the word never me ans that the soul has left the body and has become resi-
dent "in God, " as Rohde sometim es seems to suppo se. It always means that the bod y has a
god in it, just as empsychos means th at it has a psyche in it. "Deus inclusus corpore hum ano,"
Cicero wrote (De divinatione l. 67. See Dodd s 1951, 88n . 45.
5 . Phaedra , who is secretly dying because of her love for Hippol yrus, son of T heseus , her
husband .
6 . Dictynn a was worshiped in Crete, Phaedra 's own country. A deity of the heights, she
was confused with the Mothe r (L. Meridier, trans. of Hippolytus, "Les Belles Lettres ," Paris
1927, 34 ).
7 . Such as daimon, from daimonao "to be in the po wer of a god, to be possessed ," a wo rd
used by Aristophanes and Aeschylus, but one that does not occur, unless I am mistake n , in
Plato.
8 . From which we have the wo rd cateuque, whic h signifies the possessed person in Mon -
tanism . (Cf . Lacarriere 1961, 293 ).
9. Without pretending to be exhaustive and without including the German authors whom
I only know through Linfonh, I will cite Robin , Jeanmai re , Croissant, Boyance , De Mar-
tino, Hamilton and Cairn, Jowett, and Fowle r. Strangely enoug h , Dodds , who cites Linforth
several times both in Th e Greeks and the Irrational an d his com ment ary on the Bacchae,
does not mention his stud y o n Phaedrus 244d- c .
10 . Since I do not know Greek, which I can hardly spell, I owe this tran slation to Eric de
Dampierre, whose long familiarity with Plato 's language on the one hand and with ethno -
logy on the othe r has particularly well prepared him for such ente rprise. Ma y he be wa rm ly
thanked for his friendly and precious help.
f!-
/j·,
I.
11. T he division of this passage in four parts, numbered 1 to 4 is mine, and is only meant
to facilitate commentary .
12. Particularly L. Robin in his translati on of the Phaedrus (19 33, 32 n . 2) and Linforth
(1946b, 164-65 ), to quo te only these two a·n hor s, whose inte rpre tatio ns do not solve th e
problem .

..1.-. •.
No tes to Pages 194 - 205 343
13 . Cf. Boyance (1932, 65 ), who, speaking of the establishment of the Dionys us cult,
writes : "Everywhere the God is a victim of unfriendly reception, which is viewed as a real
crime of impiery. Everywhere , as punishment, a frenzy takes hold of the thiasos, femini ne
ones most of the time. And everywhere, the establishment of the Dionysus cult is the remedy
for this punishment and the indirect consequence of this fault."
14. Strangely enough, Boyance does not draw from this legend (which he reports pro- a
pos, precisely, of our Phaedrus text ) the expected conclusion. Falling in the common error,
he writes: "Plato ... sees at the origin of these purifications a madness which results from
'ancient faults,' bur which in a prophetic frenzy, knows how to disclose the liberating Gods"
(1932 , 63 - 64 ).
15. Th is has been seen very well by Jeanmaire in Le Sa tyre et la Menade (1949, 467 ).
16. Let us recall here that in the myth the first trance of Dionysus is attributed to a fright .
Let it be noted too that fright is often what causes a horse to run wild, something which might
well be viewed as an animal equivalent of human trance (see above, p. 13 ).
17. In his study entitled Le Sa tyre et la Menade (a study to which this present chapter
owes a great deal ), Jeanmaire notes that the "bacchic dances belong to an order of things
about which the philosopher [Plato], even if he disapproves of them, within himself did not
feel completely free to express himself" (p. 165 ).
18. Speaking of the evolution of Plato's ideas concerning man ia, Jeanne Croissant ( 1932,
20 ) obse rves that there existed for him a certain "hie rarch y" in the "mystic facts," and that
when Corybantism and Bacchism were involved, he was "nea rer disdain than respect."
Linforrh, ignoring the evident, defends a totally opposite position. He claims that, where
trance is concerned, he cannot find a single note of disapprobation in Plato 's writings, and is
of the opinion that they imply nothing but gratitude for some th ing admi rable, so that every -
thing should be put on the side of approbation and nothing on the other. Pushing this aberra -
tion to its limit, he also thinks ( 1946b, 171 ) that Socrates, wishing to make the reader forget
the "degrading" aspects of possession, is led to emphasize "the diviniry of Madness herself."
19 . On Corybantes and Corybantism, see in particular Jeanmaire 1951, 138££.
20 . The famous passage in the Symposium (215c - e) merel y corroborates the others . I
shall therefore not rake it into account in what follows, since it would add nothing relevant
to our present purpose .
21. The translation of this text follows the same principles as those previously adopted
for Phaedrus 244d - e.
22 . Choreia ind eed means, for Plato (Laws 654b ), both dance and singing.
23. The subject of the long sentence that follows could well be not "movement," but rath -
er "all that," meaning "all that has gone before."
24. Let it be understood once and for all that for the Greeks the word aulos designated a
cla rine t, usually a double cla rinet, with divergent pipes mostly; occasionally an oboe, but al-
wa ys a reed instrument . Never a flute, contrary to what one generally read in texts dealing
with ancient Greece, including musicological ones, even though this error has been repeated -
ly pointed out for man y years now. Such, at least, is the present state of ou r knowledge.
25. Cf. Dodds, who writes : "If the rites of a god X stimulated him [the patient] and pro -
duced a catharsis , that showed that his trouble was due to X; if he failed to react, the cause
must lie elsewhere " ( 1951, 78 - 79 ). "Hence the importance arrached to discovering the iden-
tiry of the possessing Power" (ibid., 98 ).
26. Lin forth translates (1946a ) as "words and gestures to suit thac tune."
27. On the basis of the same two passages of the Laws and Ion, Dodds (1951, 98 n. 102 )
presents the relationship between music and telestic trance in the following manner: "the
Corybantic rite included ( 1) a musical diagnosis; (2 ) a sacrifice by each patient to the god to
whose music he has responded, and observations of omens; (3 ) a dance of those whose sacri -
fices were accepted, in which the appeased deities [perhaps impersonated by the priest' ] were
 344 NOTES TO PAGES 206-14
believed to rake pa rt. " He rhus fully recognizes the role of music as a "diagnosis," rhat is ro
say as a way to identifyi ng rhe deity, bur does nor show music and dance as the wa y, for the
cory bantizer , to identi fy himself wi th the deity. Bur indeed, Plato does not do it either . .
,.i
1:1 28. Linforth, in his article on Corybanric rites (19 46a , 138-140 and 152-54 ) discusses
~: the meaning of rhis verb at lengt h. Hi s ma in purpose in doing so, howe ver, is to refute certain
theories inte rpreting the wo rd as meaning to be afflicted by the sickness of the Corybantes.
His view point is thus quite different from ours here. T hat said, for him the "fundame ntal "
meaning of the wo rd is "possession by the Corybanres." As a consequence, he transla tes it
sometimes as "possessed by rhe Corybantes" and somet imes by "taki ng part in the rites of the
Co rybantes." As for korybantiontes, he translates ir as "Corybant ic de vo tee." There is no in-
dication that for him it means "to Corybantize."
,,' 29. Regarding rhe verb baccheuein, Dodds w rites : "bacche uein is not to have a good
time, but to share in a pa rt icular rel igio us rite and (or) have a particular religious experi -
ence -t he experience of communion w ith a god which transforms a human being into a Bac-
chos or a Bacche" (1951, 271 n . 1).
30. H . Grego ire in his tr anslation of the Bacchantes (Coll. Belles Lercres , Paris 1961 ) pre -
fers another reading of this line, but conside rs ir as still "acc eptable ." In his edition of the Bac-
chae, Dodds, on rhe cont rary, does not consider the possibility of two different readi ngs. I
defer to his opinion.
31. I am referring he re , of course, to the pretheaaical form of the dithyramb, which origi-
nally must have been a sung and danced rou.id accompanied by sac rifices.
.... 32. Oreibasia: wild running about and dancing up in the mountains, considered by
(.
Dodds ( 1951, 76 and 270ff .) as being "the prototype of the ritual madness."
·.:.·· 33. "Proirus' daughters , struck by Dionysiac madness, are Argian Bacchantes, just a1
Minyas' daughters are Bacchantes from O rchomene and Dadmus 's daughters Theb an Bae·
chantes," writesjeanmaire (1951, 74 ).
34. Quoted by Jeanmaire (1951, 347 ).
35 . We here again find a topic - trance and drugs - that has alre ady been evoked , ve!)
briefly, in the first part of this book. Let us recall th at in his book entitl ed Poisons sacres, Iv-
resses divines (1936 ), Ph ilippe de Felice was probably the first to ha ve dealt at some length
wirh rhis vast problem (about which much has been pub lished since ).
36. About rhe maenad 's dancing, see the chapter "Les danses orgiast ique s" in Sechar
(1930 ), and also Lawler's study (1927 ). Since the publication of these two still classic works
other studies have o f course been pub lished; I did not ha ve the op pommi ty to consult them.
37 . Cf . Dodds 1960: "Evidence from Vase Paintings," in Bacchae xxxv - xxxvi.
38. Cf. Jeanmai re 1951, 239.
39. Cf. Jeanmaire 1951, 235.
40. Coming thi s time not from th e Greeks but from the Romans and concerning not teles
tic but erotic mania, an example of this movement of rhe coming and going of the souls i:
provided - in a rather unexpected way - by Cato the Elder who said , Plutarch reports in hi:
Erotikos "the soul of the lover lives in the body of someone else [that of rhe beloved]." Cf
Flacelier e, nores for rhe transla tion of rhe Erot ikos 1953, 130 n. 72 ).
41. Jeanmaire quotes a long passage of it in his Dionysos, p. 308.
42 . On the Platonic theory of mimesis, see J.-P. Vern ant (197 5).
43. Directly in the Symposium (215-1 6 ), The Laws (790e ), and the Crito (54d ). lndirect
ly in T he Republic (399d - e).
44. T hree in the Ion (536bc ) and one in The Laws (790 e).
45. In Th e Republic (400b ).
46. Cf. Jeanm ai re ( 1939, 579 ), quoting after Strabo, and D. W. Masaraki 1974.
47. On rhe bombyx, see D. W. Mas araki 1974.
48 . Lines 871,879, and 897.

< L. .. .
Notes t0 Pages 214- 2 8 345
49. Concerning Lyssa "T he Rage," which plays such an import ant role in the tr iggering of
"wild" trance among the Greeks, I cannot help bur think of rhe great expiatory rite pra cticed
by the vod un wo rshipers in Benin (forme rly Daho mey) and known und er th e nam e of oma, a
Fon and Gun word signifying "the rage." Cf. Adande and Verger 1953 .
50. On eher's theory, see above pp . OOOff.
51. For mo re derails on these flucruarions, see Lasserre (1954, 50- 52 ).
52. Pseudo -Plutarch, Lassere's translation (1954, 46).
53. Translated by Dodds 1960 .
54. Cf. M. Emmanuel 1914, 397 . The figures on this cup are reproduced on the cove r of
Jeanmaire's Dion ysos .
55. To be exact, nor the "melodies" bur rhe "things " or "productions" of Marsyas.
56. Let us leave aside here the problem of an approp riate rranslarion of the wo rd har-
n1onia.
57. Preserved in a fragment of a dithyramb by Melanippides the Younger. Cf. Lassere
(1954, 51).
58. Or more exactl y, since the word "melod y" here again does nor appear in the Greek,
the "things of Marsyas."
59. Olympus was the favorite pupil of Mar syas. The runes Aristotle speaks of were thus
rhe same as chose mentioned by Plato. This is made clea r, moreover, in the Symposium
(215c ).
60. Upon this point, see rhe chapter "The Physiological Bases of Musical Catharsis," in
Jeanne Croissant's book Aristote et /es Mysteres, 1932.
61. The word ethos taken in the conte xt of rhe music does in fact not form part of Plato's
vocabulary; he never speaks of the ethos of a mode. Aristotle, on rhe other hand, does so
constantly. On rhe notion of ethos seen from a comparative viewpoint, see the short accoun t
bv Curt Sachs (1943, 248,252 ).
62. One may well ask oneself, of course , if what Aristides Quintilianus calls "Plato's ha r-
monies" re ally does correspond wirh these harmonies. But rhis is another problem (Cf. I.
Henderson 19 57, 349 ).
63. Personal letter. On this point, see the article since published by S. Baud -Bovy, in 1978 .
64. Quoted by S. Baud-Bovy (1967, 20 ).
65. In Lasserre we read ( 1954:31 ): "Until the time of the sculptor Myron, and probably
until that of Plato , the Silenus Marsyas was generally regarded as the inventor of the double -
piped flute [read : aulos] ... Aristoxenes, however, taking early poetic sources as his basis ,
without doubt, subs titutes for Marsyas his father Hyagnis. But aulody itself , and sometimes
auletics, claims as its founder Olympus, the son or pupil of Marsyas, whom chronographers
put some 200 years before Homer . ." (1954 , 31).
66 . Whe ther Dionysus really did come from Phrygia or nor is of little concern to us here .
The important thing is that he was believed to have done so. On rhe very complex problem of
the o rigin of Dion ysus worship, see in particular Jeanmaire ( 1951 ).

CH..\PTER Six
1. Quotation from R . Flaceliere 1953.
2. Quotation from Jeanm aire's Di onysos (1951, 290 ).
3. Cf. Yates 194 7, 38.
4. Reference is clearly being made to Timothaeus of Miletus , a famous Greek musician
who lived in rhe fourth century B.C. , known for the boldness of his musical innovations and
mocked by Pseudo -Plutarch in his De Musica (cf. Lasserre 1954 , 145 ). Perhaps it is worth
noting in passing that if, as we are told, Timo rhaeus died in abour 357, then it is somewhat
unlikel y char he could have caused Alexande r to rise from the dinner table, since Alexande r
346 NOTES TO PAGES 229- 33

was nor born until a year later! But this is of slight importance. Our story is significant for its
symbolic content.
5. Lasserre (1955, 76 ) quotes five variants, but Jamblique gives ano ther, quoted by
Boyance (1936, 125 - 26 ).
6. Here it is the Phrygian mode that is held responsible for the young man 's frenzy, where -
as a moment before it was the H ypo phrygian. T his fact has been interpreted by M. Dabo -
Peranic (En cyclopedie de la musique, Fasquelle, s.v. "hypo phrygien" ) in a way that seems to
be without foundation. Even less wellfounded, it seems to me, is the theory defended by Las-
serre (1955, 62 - 63 ), according to which the Phrygian scale is said to ha ve been that of rea -
sonableness and moderation . Bur chis is another problem.
7. The versions of Philodemus, Galien, and Quintilian also specify it. Cf. Lasserre 1954,
76.
8. T o cell the truth, of the five versions cited by Lasserre (1954, 76 ) only Galie n's spec ifi-
cally says that the second rune, the one that calms the young Taorminian, is Do r ian. The
ochers, including Jamblique's, all speak of spondee. This spondee, Lasserre and Boyance
(who refers on this point to T. Reinach ) both say is the same thing as the "Song of Libation"
(since spondee meant libation, hence the name of the rh ythm chat usually accompanied it).
Boch also refer to Olympos who, according to Pseudo-Plutarch (De M usica, 11) invented the
enharmonic scale when playing chis song. This would mean that spondee refers here nor to a
rhythm bur to a melodic genre. But Pseudo-Plutarch indicat es that in performing this "spon -
diasm" Olympos was in the "Dorian rona!ity" (De Musica, ibid.; Lasserre 1954, 137 ). We
are thus brought back once more to the Dorian that Galien openly specifies, and this is what
we need to remember.
9. Two of the six variants, that of Galien and char of Mart. Capella, attribute the affair to
Damon, nor Pythagoras.
10. And not the "music of rhe spheres," contrary ro what is too often incorrectly stated.
11. Cf. the article by T. Reinach (1900 ) on the music of the spheres.
12. Length ratios of vibrating strings .
13. Boerhius, through whom the harmony of the heavenly bodies became known as mu-
sica mundana . The latte r, the music of the universe, formed, according to his theory, along
with musica humana and instrumental music (constiiuta instrumentis ), one of the three parts
of music seen as a whole .
14. I quote from the text reproduced by Frances A. Yates (194 7, 319 ) in her book on the
French academies of the sixteenth century, a work from which a great deal of the information
used in this chapter is taken.
15. Pierre Ronsard, Oeuvres completes, ed. Blanchemain (1966, 5:96 ).
16 . Seep. 42.
17. Let me obse rve in passing, since it is worth stressing, that Pontus de T ya rd was not
only the autho r of obscure allegories collected and published under the rather misleading title
Discours philosophiques. His repurarion as an enlightened man, and his prestige among po -
litical leade rs of the time was such rhar Henri IV, at rhe rime when he was preparing to con-
vert to Roman Catholicism, wrote to him to confide in him and ro ask his advice. Cf. Yares
(1947, 222 ).
18. The dialogue by Plato in which poetic mania is treated, as we saw in the pre vious
chapter.
19. On the reinterpret ation of Plato's four mania by Ficino, see Yates (1947, 81 ) and the
long extract she adds in a note from his commentary on the Symposium.
20. On this point, see Yates (1947, 92 ).
21 . Dodecacorde: collection of psalms composed in the twelve ecclesiastical modes, pub -
lished at La Rochelle in 1598 . Cf. H. Expert (1900 ).
22. Harmonie universe/le 1636, sig. A vi recto.
Notes t0 Pages 233 - 46 347
23. On this point, see Yates's (1947 ) cha pter, "Th e Religiou s Policy."
24. In Question es celeberrimae in Genesim, Paris 1633, col. 1532.
25 . Probabl y also the author of an ode on the effects of Claude Le Jeune's music .
26. The story does not say so , but the se two "airs" almost certa inly belong to La Gu erre,
since it was this piece, composed on the theme of love as a symbolic tournament, tha t Claude
Le Jeune wrote for these festivities. Cf. Lesure (1961, 47 ).
27. Mersenne, Correspondance, ed. P. Tannery, 1932 1 :45. Rousseau was to recount
the se sto ries in his rum, and comment o n them with great humor , in his Dictionnaire de mu -
sique (17 68, 12 ).
28 . Cf. Yates (19 47 , 2 and 39 ).
29. Preface to Melanges de cent quarante -huit chansons ... publ ished in Paris in 1572.
Cf. Oeuvres completes of Rons ard, ed. P. Blanchemain, Paris, 1866, 7:337 - 41.
30. Even thou gh things were alread y beginning to change. But this is secon dary. Th e
changes had no t yet resulted in any confusion of terminology, and this is wha t is significant
here.
31 . On this po int , see Yates (19 47, 47- 4 8) and Chailley (1956 , 138-39 ).
32 . Qu oted in H. Expert (1900 ) .
.).). It is true that Orpheus is usually depicted with a lyre. Ho wever , as Boyance (1936,
42 ) o bser ves , it is "primaril y by the power of his singing, not by th at of his lyre , wh ich ac-
companies him only after a certain date " th at he creates his effects. Acco rding to Plato (Laws
3 . 677d ), it was also, of course , to Orpheus that the invention of poetry was revealed. T he
invention of the lyre was attributed to Amphion .
34 . "Few prob lems remain as o bscure as those reg arding the relation s between Orphism
and Pythag oreani sm," Boya nce obse rves (1936, 93 ). Dodds, for his part (1951, 147ff. )
tends toward the view th at the two things are, broadl y speaking, so interconnected as to
form a single entity, and that what is generally termed Orphism cou ld ultimatel y be reduced
to Pythag orean ism Here , need less to say, it is not the mathem atical aspect of Pythagorea n-
ism that is invotl:
· ., but rather its mor al and religious aspect - inso far as it is possib le to sep-
arate the one f, the othe r. My comments on Orphic incantations apply equ ally to Py-
thagorean inca _ations.
35. Cf. Yates (19 47 , 111 ).
36 . Cf. pp. 129-33.
37 . Confessions (book 10, chap . 33 ).
38. Jeanmaire is here thinking of the pass age (Symposium 215 ) in which Plato, wit h refer -
ence to the runes of Ma rsyas, says that th ey are the only ones that put men in a state of pos -
sessio n and that it does not matter if th e autos player is a good or a bad :me .
39. Cf. R. Flaceliere's Fren ch tr anslatio n (1 953, 76 ).
40 . Linforth 1946a, 140,156, and 158 .
41. Cf. T. Reinach 1900, 433, n. 1, referring to th e Harm onia mundi ( 1619 ).
42. On this aspect of Kepler, see Gerard Simon (1979 ).
43 . Or , more pr ecisely, might have been written
44. Alada: city in southern Benin, capital of an ancient kingdom from which the king -
doms of Abomey and Porto -. ovo late r emerged. Alada is in many respe cts a veritab le con -
servato ry of ancient customs .
45. Ajah ut9: "Killer of Aja," founde r of th e dynasty.
46. Akplogan: "Chief of the lance," the principal religious figure of the kingdom, and for
this re ason often called "C hief of rhe fetishers. "
47 . Sakpa tasi : "wif e of Sakpata, " who is the god of earth and small pox . The Sakpatasi
perform particularly elaborate dances. The men execute spectacular, peri lous leaps. On the
suffix -si, see above, chap. 1, n. 74 .
48 . Legba: messenger of the gods; tr ickster and ph allic divinity.
 348 NOTES TO PAGES 246- 56

49. Khevioso: god of lightning.

50. Recade : ornamented scepte r , a veritable wo r k of arr rhar is sometim es carr ied by th e
kings of Dahomey and so metimes entru sted to messenge rs to sign ify rhe monar ch's presence .
51. For one episode from these ceremonies, cf . rhe recording by G. Rouger and P. Verger,
Dahome)\ M usique des Princes .

CHAPTER SEVEN
1. Since our examination of Sufism will rake us as far as Morocco on the one hand and
Yugoslavia on the other, in the following pages we will nor deal with only rhe Ar abs. Should I
have used, in this case, the term Islam ro cover this vast field' T he manifestations with which
we will be concerned are occasionally so far from Islamic orthodoxy that using the wo rd Is-
lam would still nor have solved the problem. In any case, rhe Arabs form such a significan t
pan of the chapter that iris only logical they should figure in irs title.
2 . Sufism: the wo rd has been defined (T rimingham 1971, 1) as denoting that mystic ren-
d ency in Islam char aims "a t direct communion benveen God and man ." T he definition is
schematic bur convenient. Among rhe various erymologies of the word advanced by the
Arabs, we should rec all the one that identifies the rerm Sufism as deriving from the word suf,
a woollen garment habirually worn by Moslem ascetics. The Sufis are organized into a large
number of brotherhoods ('tif'i/a ), each of which as its ow n "way" (tariqa ). For a general
overview of Sufism, cf. Corbin ( 1964, 262ff. ).
3. Modern life has without doubt made such behavior much rarer, bur it still occurs quire
commonly in va rious regions, as. to cite only one example, in rhe sout h and central region of
Iraq (S. Q. Hassan, personal communication ).
4 . An English translation of the Kiffbadab al-sama-<wa al-wajd was published in 1901 - 2
by Duncan B. MacDonald in an article entitled "Emotion al Religion is Islam as Affected by
Music and Singing." This is the version which I have consul red and to which I refer in the rexr
under !ac Donald 's name. The last pages of the Kitiib also appear in French translation in
M. Mole's long article (1963, 193 - 201 ) entitled "La danse exrarique en Islam."
.....,r",_.1rh ough Ghazzali is highlighted in this chapter, this does nor mean that his work is rhe
~ one of irs kind, or even rhe most impo rt ant . Man y other Sufis had dealt wirh rhe prob-
lem before him, among rhem, to cite only one, al-Hudjwiri, whose Kach/ al-mahdjub ("Un-
veiling of Sacred Things ") has been translated into English by Nicholson (1911 ). And orhers
\\·ere to deal with it afte r Ghazzali. Bur for our purpose, which was nor ro enter into the de-
rails of the arguments aroused by the great question of music's "lawfulness," it seemed berre r
to focus attention on a single author, who is in any case quire representative of rhe Sufi pro -
music tendency.
6 . In his rrans lar ion of the Kitab, M acDonald rranslares samr sometimes by "Listen ing to
Music and Singing," sometimes by "He aring Music and Singing," sometimes by "Listening,"
sometimes by "Hearing," sometimes by "Music and Singing"; a rather unsatisfying diversit y.
His foomore ( 1901, 236 ) on tran sla ting the term makes his uneasiness quire clear.
Robson (1938 ) in his English transl at ion of rhe treatise by Ghazzali's brother Ahmad on
samif, rranslares the words as "audition." Farmer generallr translates ir as "Listening to Mu -
sic," bur he (19 29, 140 ) also someti mes rranslares ir simply as "music .··
In French, L. Massignon (19 22, 8Sj rranslares samii' as concert spirituel or "o ratorio,"
terms later raken up by \1. Mokri (1961 ) and E . de Vir ray- Meyerovirch ( 1972 ). Tri -
mingham ( 1971 ) also translates it as "spiritual concert." H. Corbin (1964, 260 ) translates it
as ·' audition musicale .., Mole, in his long article on ecstatic dance in Islam , uses rhe word sa-
mif throughout without attempting ro translate it. As we can see, this is not an eas y word to
transl are.
7 . The question of whether music in general is o r is nor lawful in the eyes of rhe faith, and
whethe r it should be permitted or forbidden, has preoccupied Muslims from rhe very srarr,
as is "·ell known. And rhis applies e,·en more ro the Sufi samif , since its music is in direct rela-
 ores co Pages 25 6- 63 349
rion to religious pr actice. On this ext rem ely muddl ed issue, f.ee w h at I feel to be a very good
summary made by Farmer (1929, 20-38 ) in his A H istory of Arabian Music, in rhe chapter
"Islam and Music."
8. The Ar abs have written a great deal abou t sama-<,either to praise o r condem n it. A cur-
sory inventory, based on MacDonald (190 8 ), Farmer (1929 ), Robson (1938), Mokri (1961 ),
and Mole ( 1963 ) of the works entirely or p artly devoted to sama-<,provides us w ith a list of
twenty -six works, seventeen of which have the word san1t1-cin rheir rirle. One of rhe oldes t is
that by al-Hujwiri (d. A.D. 1072 ), from which Ghazzali makes numerous borrowings . Th e
Kitab sharh al -sama-c"Commentary on the Sama-<") by al-Farabi (no. 165 in Farme r's bibliog-
raphy, 1940 ), which is much earlier, since the autho r died in A.D. 950, has unforrunarely been
lost.
9 . Mole (1963, 148, 227,235, 237 ), quoting var ious Arab autho rs .
lC. Mole (1963, 204 ) quoting rwo different Arab au thor s.
11. In his treatise on sama-c, Ahmad, Ghazzali's brother, uses the word musiqi once (Rob-
son 1938, 71- 122 ), in a very gene ral sense.
12 . Malahi, meaning "ins truments of entertainme nt," in other words "mus ical instru-
ments" but with a moral nuance in the meaning, derives from lahw.
13. For the term taghbzr, see Farmer 1929, 33 n . 2.
14. On rhe me ani ng of wajd, see Garder (1976, 17 5- 83 ) and his rem arks on Massig n on's
interpretation of the wo rd.
15 . l acDonald translates it both rimes as "revelation." I am told by R. P. Nwyia rhar rhe
word "inspiration " is nearer to the Arabic .
16. A lure with a long neck.
17. From a sixteenth -century rexr quoted by Rouaner ( 1922, 698 ).
18. Kitab al-aghan i. Tr anslation by Quatreme re, Journal asiatique 16 (1935 ):51 5 and
517.
19. Farmer is here quoting from Isfahani's Book of Songs.
20. Al-Junai'd : master of Baghdad Sufism , acclaimed as spiritual ancestor by the majority
of Sufi cong reg ations (Mole 1963, 169, n. ).
21. We will deal with dhikr a lirrle further on .
22. A famous Sufi mystic who lived in Baghdad in our tenth century and claimed unity
wirh God, was declared a heretic and crucified.
23 . By Ghazzali notably, in another book of his great work on rhe revivification of faith.
For a bibliography on dhikr, see Garder (1952 - 53 or 1976 ) and Trim ingham ( 1971 ).
24. To do so , 1 shall mainly use rhe two studies by L. Garder ( 1952 and 1953 ), "La men -
tion du nom di,·in dans la mystique musulmane." These studies have been edired and pub-
lished in book form in 1976 in collaboration with G. C. Anawari.
25. On dhikr technique, both in itself and compared with orher types of mystic orison
("Jesus prayer" of rhe hesychasrs and rhe nembutsu of Japanese Buddhisrs in particular ), as
well as its relations wirh yoga, See Garder 1952 - 53.
26. Dhikr or zikr: orthographical variants rhar indicate variations in pronunciation.
27. \Y/eknow chat the Arabic f aqir (Persian darwish ), which means "poor,' ' designates an
adept among the Sufi. Ir is therefore logical to speak here of fakirism . However, in both
French and English the word usually has a pejorative connorarion and denotes activ ities usu -
all y regarded as more o r less contemptible and fraudulent. Need I say char my use of it here
do es nor include any such nuance. The facr that rhis particular aspect of fakirs' activities,
which consists in public performance of exercises aimed at amazing onlookers because the y
challenge (or seem to challenge ) certain laws of nature - Massignon (1934 , 701 ) refers to
them as jugglery - has been severely criticized by canonical Muslims, by most Christians,
and by a good number of nonbelievers, has of course nor been taken into consideration here .
In Morocco, where the word dervish is nor used, they say f agir (pl.fuggara ).
28. The Turki sh form is Mevlevi.
350 NOTES TO PAGES264-73
29. We shall return to this word in a moment .
30. T he Yezidi form a "heterodox Islamic sect" (H assan 197 5,234 ) an d are no t Sufis. See
also Hassan 1976.
31. Ibn Jub air , a tw elfth -century travele r , writing in rhe rime of Sala din , describes with
wonde r the "impa ssive sama'" during whi ch the participants become so "enraptured" in ec-
srasy "that the y can hardl y be rega rd ed as belongi ng to this wo rld at all" (from T rimingham
1971,10 ).
32. Cf. Mole 1963 .
33. Quoted by Fanner 1929, 140.
34. Cf. Mole 1963, 238ff.
35. Here we once again encounter the word /arab, albeit in another form ; we shall return
to it later.
36 . Is the camel pa rt icularly sensitive to music' One would be tempted to think so upon
reading rhis text, particularly when it is compa red to another, by Jalal al-Din Rum1rhi s time,
who in his Mathnaw i mentions a shephe rd who played the flute so well that "rhe camels
themsel ves made a circl e around him" (quoted in E . de Virray -Meye rovit ch 197 2, 88 ).
Mo reover, in T urkey, we know that they hold (or still held recentl y) camel fights in whi ch the
rwo animals are incited to combat by two bagpipes played by the respecti ve camelherds , each
of whom is pressed against rhe animal's flanks with the intention of driving him into a frenzy.
(I learned this information from someone who observed such fights himself, and w ho al-
lowed me to listen to his recordings. )
37. Cf. Mauguin (discography ).
38 . From E va de Virray -Me yero ,·irch (197 2 , 83 ).
39. A fact rhar emerges, at any rate, from descriptions of the Qadiri yya and the Rif'aCiyya
by Depont and Coppolani (1887 ); the ' Isawiyya in Morocco and Algeria by Brunel (1928 )
and by Dermenghem and Barbes (1951 ); the Rahmani yya of Alge ri a by Ha as (1943 ), quoted
by Garder (1952 ); rhe Shadhil iyya of northern Africa and rhe Mirghabiyya of rhe Sudan by
Trimingham (1971 ); the Sadu liyya of Tuni sia by Ferchiou (1972 ); the Rifa<iyya in Syri a and
Yugoslavia by (respecti vely) Poche (1973, discogr aph y) and Mauguin (1974, discography ).
T he performance given in Paris in 1978 by the Khalwati yya (or H alveri) from T urkey fol -
lowed the same scenario and enabled me to grasp cert ain aspects firsthand, thus pro vidin g a
very useful supplemen t to my reading .
40. Hiz b mean s "part , section" (of the Koran ). The word wird is also ofte n used to denote
all that is recited or sung du ring this first part.
41. Ir is worth noting rhar these "so ngs " (T rimingham 's rerm ) are nor design a red by rhe
wo rd ghin a', which is the term usually emplo yed in Arabic for "so ng" and has almost rhe
same meaning as ou r word "music. "
42. R. P. Nwyia w rite s me, in a lette r , rhar this is probably the first sura (h) of the Koran,
the title of which means "that which opens ."
43. The se singers are not necessarily members of the brotherhood. Ar lea st this is nor the
case among rhe Rifa'iyya of Aleppo (cf. Poch, discography ). Moreover the presence of such
singe rs does not always seem to be required. Ferchi ou (1972 ) makes no mention of them in
her description of rhe dhikr of the Tunisian Saduff yya .
44. Female Sufi colleges do exist (T rimingham 1971, 18). Do they practice the dhikr?
One is tempted to assume so . But perhaps it is not practiced in public, which would expl ain
why there is no description of chem, to my knowledge at least.
45. Cf. T rimingham (1971, 197 and 210 ).
46. On this point, see Garde r (1952, 659 ).
47. Both of which have spread out a gre at deal. Cf. Trimingham (1971, 37- 40 and 86 ).
On the expansion of rhe 'Isawiyya, see Brunel ( 1926, 46ff .).
48. Called darb shish, "' piercing ' with the swo rd" (Poche, 1976, discography ). Th e
skewer made of metal with an ornamented cha in, is about fifty centimeters long (Poche, per -
sonal communication ).
Notes co Pages 2 7 4 - 8 2
351
49. Voyages d'Ibn Batoutah , quotation in Mole (1963, 223 ).
50. _ T he author refers ~nl~ ro_north Africa, bur it seems likely th at the ceremon y took
place m Algena since the A1ssaoua of Morocco do nor practice this kind of ordeal, according
ro Brunel.
51. Al-RifjiCi'sfirst name was Ahmad, hence this othe r name for the Rifa"iyya.
52. Putting the Rifa"iyya and Anasren arides rogerher is quire legitimate: Sufism and East -
ern Christianity are not at all foreign ro one anothe r.
53. On th e Anas renarides and their ordeal by hot coals, see Kako uri (1965 , 7- 32 ).
54. Quotation in Mole (1963, 223 ).
55. T rimingh am writes ( 1971, 86 ) that the ' Isawi yya borrowed these "ecsta tic practices,
where by the dervishes became immune ro swo rd and fire", from the Rifa"iyya.
56 . Or at any rare we re practiced in this way in Morocco in 1926. All the informa tion
used here is taken from Brunel (1926 , 92-122 ). The ' Isawiyya are also called ' Ai:ssaoua.
57 . On this point see Brunel (1926 , 233ff. ).
58. Seep. 000 .
59. T he ' Isawiyya pr actice a rite involving the earing of raw flesh, rhe/ rissa (Brunel 1926 ,
17 5ff .) rhar has often been compared with the omophagy of the Bacchantes .
60 . Brunel makes brief allusions ro them on various occasions.
61 . The one whos e music was recorded and published by Poche, mentioned several rimes
in this chapte r, provides a very abridged bur very significant example, with rhe one exception
that the wird - rhe introductory part - is omitted. T he particular fearure of rheh,a,jra record -
ed by Poche is rhar the skewe r ordeal precedes rhe dhikr proper, whe reas the reverse order is
usually rhe rule: the faki risr pra ctices normall y follow the dhikr. This is the order followed,
among other examples, in the seanc e described by lbn Barurah .
62. Acco rding ro MadDonald and Masse (Encyclopedie de /'Islam 1965 , s.v. "Jinn" ) rhe
possibility rhar rhis wo rd was borrowed from rhe Larin genius cannot absolu tely be ruled
our.
63. T he T idjaniyya congregation displays very differen. fearures in Morocco and Sen-
egal. On possession among th ese T idjaniyya women from Tunisia and its relation ro music
as they emerge from Ferchiou 's desc ription, see p. 000 .
64. For a definitio n of jinn, see the art icle by MacDonald and Masse cited in note 62
above.
65 . On possession by jinn cf. Encyclopeadia of Relig ions and Eth ics, 1918, vol. 10, s.v.
"possession," pp. 135 - 36 .
66. Taken from Farmer 1929 , who himself refer s ro Hir schfeld (1902).
67 . Cf . p. 000 .
68 . "In order ro go into ecstasy, he liked ro set a rose, as a sign of separation, between
himself and the face of a young novice " (Massignon 1922, 88 ). On eroticism in samac and its
profound ambiguity, see Mole (1963, 155 - 56 ).
69. On Jamila, see Farmer 1929, 85 .
7C. On rhe word tarab, in the context of music, see Shiloah (1972, appendix 2 ) who
writes in particular : " Tarab compr ises quite a scale of affective categories: pleasure, delight ,
mental pastime, emotional shock, ravishment, exalta tion, and ecstasy, which sometimes
brings about the death of the listener." The fearures shared by wajd and rarab are indicated in
another wo rk by Shiloah (1967 , 192 n. 6 ). On !arab, see also H abib Hassan Touma (1976,
52 ).
71. These potte ry drums, are quire inex pensi ve , and the musicians are careful ro bu y "sev-
eral at a rime"!
72. Cf pp.267 - 68 .
73 . Maqam singe rs and instrumentalists are nor called mutrib (S. Q. H assa n, personal
communica tion ).
74 . Contra: Farmer (1959, 2 ) writes that in the tenth century al-mutrib was the usual
word for "musician," particularly if he was an instrumentalist. In Turke v (Mauguin, discog -
i
I
352

their spinning dance is called mitrib.

NOTES TO PAGES 283- 95

raphy ) the group of instrumentalists and singers that accompani es the Mevlevi dervishes in

75. Translated by Farmer (1959, 2). The word mutrib can thus deno te sometimes music ,
and sometimes the musician.
76. At least this is what emerges from the description given by Ga rdet (19 52 , 559 - 60 ) and
derived from Ha as (1943 ). To be sure, however, one would have to read Haas himself , which
I ha ve unfortunately been unable to do.
77. To judge from their perform ance in Paris in 1978.
78. Mauguin (discography ) tells us nothing on this point about those in Kosovo, Yugosla-
via.
79. Cf. the commentary to Poche 's record.
80 . Cf. pp. 132- 33 .
81. Cert ain very limited aspects of the problem were treated in the first part of this book:
the case of Dav id and Saul, that of rhe possessed nu,s of Loudun , and that of tarantism.
82. On "e nthusiasm" in Christi anity, see Knox (1 950 ).
83. For the first see the nineteenth -centur y engra ving depicting a Shaker dance held in
their parish house in New Lebanon (Les Shakers 1976 ); for the second, see the descriptions
of P. de Felice (1947 ) and Hem (1928 ).
84. I am spe aking from very brief person al experience in H arlem and, above all , from
what I have been told by friend s with more experience of these thing s. Igno rance or real ab -
sence of evidence? I have not been ab le to con su lt descriptions of Baptist or Pentecostal wo r-
ship that would elucidate the points of interest to us he re.
85. For a summary of the "Welsh Revival," see P. de Felice ( 1947 , 273ff. ).
86. Cf. also P. de Felice (ibid., 193££.).
87. In the Albertina Museum in Vienn a . Reproduced by F. Lesure ( 1966 ) in his illustrated
wo r k on music and society .
88 . T o speak of "Arab" phil oso phers or musicologists, to be sure , means I am reverting to
an obsolete and justly criticized usage (cf. Corbin 1964 , 5ff. ), bur from our present stand -
point it has no important drawbacks and has the virtue of simplicity.
89 . Cf. Cowl (1966, 137 and 14 8).
90. Since I h ave not checked the Arabic text I cannot assert this with cert ainty, but it
seems likely rhar rhe word used by al-Kin d, is ta'thi r or "influence, " a term that should be re-
gard ed , in this context, as equ i\'ale nt to the Greek ethos (Farmer 1929, 76 ). It is the word
th at , acco rding to Farmer (194 0, nos. 25 and 192 ), figures in the Book of Musical Cycles in
which it is transl ated as ethos by Erlanger (1938:43 ), and also in the Epistl e of the Brothers
of Purit y.
On the ide as relating to the "influen ce" (ta'thi r) of music among the Arabs and what the y
owe to ancient Greece and Mesopotamia, see Farmer (19 26 and 1957 ).
9 l. On this aspec t of al-Kin dI's wo rk , see Farm er (1926, 14ff. ) and Chattin (1939, 83 and
t:
,, 121 ),
92. It figures in the Commentaires on Safi al-Din (Erl anger 1938, 47 ).
93. Cf. above, note 8 .
94. Denouncing the "vag ue and outdated philosophy " of "those who were unable ro rec-
ognize the true nature of each scienc e," Avicenna wrote, clearly with the Pythagoreans in
mind : "l shall nor waste time explaining the principles of the science of numbe rs. . . Nor
shall I seek to establish a relati onship benveen the stares of the heavens, the cha racter of the
soul, and musical inter\'als'' (Erlanger 1935, 106 ). On Avicenna's ideas about the relati ons
between modes and the va riou s periods of the day, see Farmer (1926, 24 ), w ho refers to a
text th at does not appea r in Erl anger.
95 . Tran sla ted and edited by Shiloah (19 72 ).
96. Shiloah (1972, 45 and 46 ).
97. Translated and anno tated by A. Shiloah (1965 and 1967 ).
Notes co Pages 295 - 3 10 353
98. In the summary they give of thei r lett ers at the end of the collection. The text is taken
from Erlanger (1930, xxvii ).
99. Shiloah 1967, 192 - 93 . T he wo rd tr anslat ed here as "mu sic" is neither musiqi nor
ghinac but angham, which means "melodic mod es" (Shiloa h 196 7 , 192 n . 1).
100. Dictionnair e de la musique, s. v. "Musique ."
101. The rending of garments ( tamzzq ), the curious prac tice that amon g th e Ar abs can be
a sign either of profane or religious trance, should undo ubtedly be co mp ared with that other
form of trance behavio r manifested by the Old Testament He br ew pr oph ets (such as Saul, to
take only one example ) and also by cenain possessed pe rsons amo ng the Ni ger Songha y
(Rou,ch: persona l communication ) wh ich consists not in tea ring one's clo thin g bu t simpl y in
remo ving it. In other wo rds in stripping oneself naked, whic h is a radical and ex peditious
way of changing pe rsonality. Since both practices consist in a negation of clothin g , perhaps
they are both part of one and the same type of behavior as far as personal ity stru cture is con-
cerned. If so, then tamziq ma y well have some link to possession, but this still ne eds to be
proven.
102. Cf. pp. 167 - 68.
103. This idea is not a new one. Some two thousand years ago, Dem ocritus already ob -
served that "the act of procreation is similar to a small fit of epileps y." Apparently , at least a
fit of epilepsy and a fit of trance may sometimes be likened. Delatte ( 1934, 66 - 67 ), to whom
I owe this reference to Democritus, considers that for the latter, epilepsy, "enthusiasm, " and
orgasm fall under the same order of physiological phenomena.
104. Cf. the notes for the recording he made of the Rifa"iyya of Yugoslavia already men -
tioned. The texts in question (Mauguin, personal communication ) are being translated.
105. In fact, Garder, in his srudy, is mainly concerned with the solitary dhikr, which is
practiced in immobiliry, or almos t , but is nevertheless an exercise centered on a certain
breath-control technique. If what he says is ,·alid for this almost incorporeal form of dhikr,
then his physiological interpretation is even more applicable to the cc llect ive dhikr, in which
the entire body is involved, and very violently at that.
106. Cf. pp. 109 - 111 .
107. Borrowing the opinion of the famous Sufi marryr Hallaj, and repe ating it as if it were
his own, Massignon (1922, 86 ) expresses the most severe disapproval of the dhikr that is
practiced with the aim of triggering ecstas y "by force and through a mechan ical process as it
we re .'· Garder also writes , "any attempt to produce trance automatically . is scorned by true
Sufis" (1952 , 662 ).
108 . Cf. pp . 71 - 72.
109. This observation does not appear in the first version of the description, published in
T r imingham's Islam in the Sudan ( 1949 ). When he repeated it in his T he Sufi Orders in Islam
( 1971 ), he did so verbatim with the exception o f one short passage. The words: "T he recita -
tion and movements naturally control the breathing - an important thing if emotional effects
are to be produced, " are now replaced by the observation I provided in my text. In other
words, it \\·as not rill later on that the author became aware of the paradox of this dhikr prac -
ticed without any thought of producing trance.
11 0 . During a performance given in Paris ( 1978 ) by a H alve ti group, one of a number
they gave during a theatrical tour of France. Dubious evidence, one may say. Not necessarily,
since dhikr seances are by no means incompatible - very far from it- with the presence of an
audience.
111. Fo rmerl y Mazagran . I owe this information on the Derqawa to B. Halff.
112. Thi s information, whi ch I owe to Scheherazade Q. Hassan , applies to the dhikr in
Baghdad, but it certainly is generalizable .
113 . Cf. pp. 273 .
114 . Bands 1, 2, and 3 on side one of the record (Poche 19 76, discograph y).
115. Brunel(1926, 103 ).
354 NOTES TO PAGES 310- 22

116. Dermenghem and Barbes (1951, 309 ). Guegba or qasaba, Arabic word denoting the
transverse flute called nay in Persian.
117. The Gnawiyya, or Gnawa, form another brotherhood, very widespread in Moroc-
co, in which blacks predominate.
118 . Hal: trance State. Cf. p. 268.
119. Cf . pp. 96 - 102.
120. Brunel (1926, 98 ).
121. Ibid. , 103 .
122. Ibid ., 225.
123. Ibid ., 119.
124. Ibid ., 122.
125. On the ahwach, see Chattin (1939, 23ff. ) and B. Lortat -Jacob (discography ).
126. Term also used by Massignon , (1922, 87 ), who translates raqs by ecstatic "dance" of
jubilatio n.
127. In his description of the ahwach, Chattin ( 1939, 25 ) calls attent ion to the fact that its
"movements are very supple and discreet, and do nor in any way resemb le the brutal disloca-
tion observed . . . in the course of the ecstatic dances of the Aisawa .... " If brutal, the
dance is ecstatic, if supple, it is nor? There seems to be reasonable ground for doubt here,
since the distinction appears to be quite subjective . Besides, from one dance to the other
things change quite significantly depending on circumstances. This being so, Chottin's re-
marks certainl y confirm that the comparison made here between ecstatic dance and ahwach
is justified.
128. About the ritual music of the ' Issawiyya of Morocco, see Andre Boncourr's work
(1980 ) which appeared as this book was already in press and could nor be cited here as it
should .
129. Cf. pp. 77 - 78.
130. Lapassade interprets the Gnawa's possession rituals as falling into the realm of exor-
cism. His interpretation is thus subject to the reserv ations made earlier in the section "Exo r-
cism>" in chapter 3.
131. Cf. Castaldi (1976, 58 ).
132. Ibid.

CHAPTER EIGHT
1. I owe this information to Steve Feld and Hugo Zer,ip, respectively.
2. Cf. 1960 edition, p. 336.
3. Contrary to what one might expect from its ride, Erika Bourguignon's article (1965 ),
"Th e Self, the Behavioral Environment, and the The ory of Spirit Possession" does nor in fact
constitute a "theory of possession."
 Bibliography

The bibliogra phy, d iscogra ph y, and filmog r aph y inclu de only those wo rk s men-
tioned in the cou rse of th is book.

Adande, Alexand re, and Pierre Verger. 1953 . Un rite expi ato ire : Om a. Notes afri -
caines 58: 41 - 46 . D aka r: Institut Franc;ais d'Afr ique o ire.
Adler, Alfred, an d Andras Z emplen i. 1972. L e baton de /'aveug/e. Paris: He r-
m ann.
Akindele, A., and C. Aguessy. 1953 . Contribution a l'etude de l'histoire de /'an-
cien royaume de Porto -Novo (Memo ire de l'Institut Franc;ais d'Afriq ue No ire ,
no . 25 ). Dakar: !FAN.
Anawati, G . C., and Louis Ga rdet. 1976 . M ystiq ue musulmane . Aspects et ten -
dances - experiences et techniques (Etud es musulmanes 8). Paris: J. Vrin .
Anon ymous. 1976 . Les Shakers (Ca talogue de !'exposition , "Les Shake rs"). Paris:
CCI Edi tion .
Arbm an, Ernst , 1963 - 70. Ecstas y or religious trance. Edited by Ake H ultkrant z.
3 volu mes. Uppsala: Svenska Bokforlaget.
Aristotle. The Politics .
Arom , Simkha, and Gen evieve T au relle. 1968. Cul te des jum ea ux chez Jes Ali .
Riruel de guerison chez !es Ngbaka -M an djia en Republique Centr africaine . En -
cyclopedie des musiques sacrees 1: 92- 104 . Paris: Laberge rie.
Aubin, Dr. H. 194 8 . D anse myst ique, possession, ps ychopathologie. L'Evolution
ps ychiatrique 4 :191 - 215.
Augustine, Saint. 1926. Confessions. Text establ ished and translated by Pierr e de
Lab riolle. Paris: Societe d'edition Les Belles Leme s.
Avicenna. See Erl ange r 1955.
Bake , Arnold. 1957 . T he music of India. In T he new Oxfo rd history of music
1:195 - 22 7 . London: Oxford University Press.
Barton, George A . 1918. Semitic -?nd Christian possession. In En cyclopae dia of
Religions and Ethics , 10 :133 - 39 .
Basile de Cesa ree, Saint. 1857 . Opera omn ia . .. Edi ted by par J.P. M igne . Par-
is.
Bastide, Roger. 1955 . Le principe de coupu re et le componement afro -bresilien.
Ana is do XXXI congr . internacional de americanistas, pp. 493 - 503 . Sao Paulo .
_ _ _ . 1958. Le ca ndombl e de Bahia (Rite nago ). Paris : Mouton.
_ _ _ . 1972. Le reve, la transe et lafolie. Paris: Flamma rion .
Baud-Bovy, Samuel. 1967. L'acco rd de la lyre antique et la musique populaire de la
Grece moderne. Revu e de Musicologi e 53 :3- 20.
356 BIBLIOGRAPHY

__ _ . 1978. Echelles anhemitoniques et echelles diatoniques clans la chanson

grecque . Schweizer Beitrdge zur M usikwissenschaft, 3d ser. 3: 183-201.
___ . 1978. Le dorien etait-il un mode pentatonique? Revue de Musicologie
64:153 - 80.
Beattie, John. 1969. Spirit mediumship in Bunyoro. In John Beattie and John
Middleron, editors, Spirit mediumship and society in Africa, pp. 159 - 70. Lon-
don: Routledge and Kegan Paul.
Beatt ie, John, and Joh n Middleton, editors. 1969. Spirit mediumship and society
in Africa. London: Routledge and Kegan Paul.
Belo, Jane. 1960. Trance in Bali . New York: Columbia University Press.
Berliner , Paul. 1975 - 76. Music and spirit possession at a Shona bira. African Mu -
sic 5:130 - 39.
Bible. 1939. The complete Bible: An American translation. Translated by J. M .
Powis Smith. Chicago : The University of Chicago Press.
Binon, Gisele. See Cassard.
Blacking, John. 1973. H ow musical is man? Seattle and London: University of
Washington Press.
Boethius. 1867. De institutione musica. Edited by Godofredus Friedlein . Leip zig.
Bancourt, Andre. 1980. Rituel et musique chez /es Aissaoua citadins du Maroc,
Universire de Strasbourg. 2 vols. roneotypes.
Bourguignon, Erika. 1965. Th e self, the behavioral environment, and the theory
of spirit possession. In Melford E. Spiro, editor, Context and meaning in cul -
tural anthropology, pp . 39 - 60 . ew York: The Free Press .
___ . 1968. World distribution and patterns of possession stares. In Ra ymond
Prince, editor, Trance and possession states, pp. 3-34 . Montreal.
___ . 1973. Introduction: A framework for the comparative study of altered
states of consciousness. In E . Bourguignon, editor, Religion, altered states of con -
sciousness and social change, pp. 3-3 8. Columbus: Ohio State Unive rsity Press .
___ . 1974. Cross -cultural perspectives of the religious uses of alte red states of
consciousness. In Irving I. Zaretsk y and Mark P. Leone, edirors Religious move-
ments in contemporary America, pp. 228 - 43. Princeton: Princeton University
Press.
___ . 1976. Possession. San Francisco: Chandle r and Sharp.
Boyance, Pierre . 1936. Le culte des muses chez /es philosophes grecs (Bibliotheque
des Ecoles franc,aises d'Athenes er de Rome ). Paris: Editions E. de Boccard.
___ . 1946. Les muses et l'harmonie des spheres. In Melanges dedies a la me -
moire de Felix Grat, pp. 2-1 6. Paris.
Brandil y, Monique. 1967. Un exorcisme musical chez les Koroko. In T.
ikiprowetsky, ediror, La musique dans la vie, pp . 31- 75. Paris: O.C.O.R.A.
Brosse, Therese. 1963. Etudes instrumentales des techniques du yoga. Public a-
tions de !'Ecole franc,aise d'Exrreme -Orient, vol. 62.
Brunel, Rene. 19 26. Essai sur la confrerie religieuse des ' Ai'ssaoua au M aroc. Par -
is: Paul Geurhner.
Bruno, Jean. 1973. Extase, transe, experimentation. Critique 312:418 - 46.
Butt, Audrey . 1962. Realite et ideal clans la prarique chamanique. L'H omme: Re -
vue franqaise d'anthropologie 2:5 - 52.
Bibliography
357
Calame -Gri aule, Genevieve . 1965 . Ethnologie et langage . La parole chez /es
Dagon. Paris: Gallimard .
___ . 196 8. Dictionnaire Dagon. Dialecte taro . Langu e et civilisation. Paris: C.
Klincksick .
Carape tya n, Armen . 1948. Music and medi cine in the Renaissance and in the 17th
and 18th Centuries. In D. M. Schullian an d M. Schoen, editors, Music and
medicine, pp. 117 - 57. New York: H enry Schuman.
Carpitella, Diego . 1966. L'exorcisme choregrapho -mu sical du Taren tisme. In E.
De Manino , La terre du remords, pp. 355 - 65 . Paris: Gallimard.
___ . 1967. La musica nei rituali dell'argia. In C. Gallini, I rituali dell'arg ia.
Cassin, Elena. 1962. Review of E. De Man ino, La terra de/ rimorso ...
L'Homme: R evue franqaise d'anthropologie 2: 131 - 33.
Castaldi, Paolo, A. Pesce iri , and A. D'Angelo. 1976. La relig ion e e il pop dei
popoli, Muzak (?):5 8- 59. Rome ( ' ).
Ceneau, Michel de . 19 70 . La possession de Loudun . Paris: Jullia rd.
Chailley, J acq ues. 1956 . Le myrhe des modes grecs, Act a Musicologica 28, fasc.
4:137-63.
___ . 1960 . L'imb roglio des modes. Paris : A . Leduc.
Chapman, Anne. 1972. Selk'nam (Ona ) chants of Tierra de/ Fuego, Argentina : 47
Shaman Chants and Laments. Des cript ive notes. New York: Folkways Records
(cf. discog raph y).
Cheron , Georges. 1931. Le dyide. Journal de la Societe des Africanistes 1, fasc.
2:285 - 89.
Chattin, Alexis. 1939. Tabl eau de la musique marocaine. Paris: Paul Geurh ner.
Colson, Eli sabeth . 1969 . Spiri t possession among the Ton ga of Zambia. In John
Beattie and John Middleton, editors, Spirit mediumship and society in Africa,
pp. 69 - 103. London: Routledge and Kegan Paul.
Combarieu, Jules. 1909. La musique et la magie . Paris: Picard.
Condominas, Georges. 1973. Pos rface pour l'Cha manisme er possession." Asiedu
Sud -Est et Monde lnsulindi en , Bulletin du CeDRASEM I, 4, no . 3. Paris:
C. .R.S. er E.P.H.E . (6e section ).
___ . 1976. Quelques aspects du chamanisme er des cul res de posse ssion en Asie
du Sud -E st er clans le monde insulindien. In L'autre et l'ailleurs . H ommage a
Roger Bastid e. Presente par]. Poirier et F. Raveau, pp. 215 - 32. Paris: Berger-
Levraulr.
Corbin, H enr y. 1964 . Hist oire de la philosophie islamique, vol. 1. Paris: Galli-
mard .
Cossard, Gisele (Gisele Binon-Cossard ). 1967 . Musique clans le Candomble. In T.
Tikiprowersky, editor, La musique dans la vie, pp . 159-207. Paris:
O.C.O.R .A.
___ . 1970. Contribution a l'etude des candombles au Bresil. Le candomble an -
g o la.- Paris: Faculre des Lerrres er Sciences humaines de Paris, 2 vols . roneo-
rypes.
Courlander, H arol d. 1944 . Dance and dance -drama in H aiti . In Franz iska Boas,
editor, The Function of Dance in Huma n Society, pp. 35 - 45. New York : The
Boas School.
 11;:,
•JJ1I"
IIJ
1·
!I j 358 BIBLIOGRAPHY
ti~ l
I·
; .
Cowl, Carl. 1966. T he Risal a fi khubr ta' lif al-'a lJ:ian of Ja ' qub ibn IsJ:iaq al-Kindi
'I
(790 - 87 4 ). T he consort journal of the Do /metsch Foundation 23: 130-66 .
Croiss ant, Je an ne. 1932. Aristote et !es mysteres (Bib . Fae. Philosophie et Lettres
de Liege). Liege and Paris: Dro z.
Danielou, Alain. 1967 . Semantique musicale. Essai de psychophys iologie audi -
tive . Pari s: Herm ann.
___ . 1975. Musiques et danses d 'exrase. Le Courrier, October 1975 . Paris:
Unesco.
Defradas, Jean . 1968. La divina tion en Grece. In A. Ca quot and M. Leibov ici, edi-
tors, La divination, vol. 1, pp . 157 - 195. Paris: P.U .F.
(.
Delab y, Laurence. 1976. Chamanes toungouses (Erudes mongoles .. . et siber-
iennes, cahier 7). Nanterre: Universite de Paris X.
___ . 1977. Routes et chemins d 'esprit chez !es Tou ngouses. L'Et hnographie
n.s. 74 - 75:189 - 195.
Del atte, A . 1934. Les conceptions de l'enthousiasme chez /es philosophes preso -
cratiques (Extract from L'Antiquite classique, vol. 3) "Col lection d'e'rudes an -
ciennes. " Societe d'edition "Les Belles Lettres." Paris: Dro z.
De Martino, Ern esto. 1966. La terre du remords. Tr ansla ted by C. Poncet. Paris:
Gallimard.
Densmore, Frances. 1948. The use of music in the tre atment of the sick by Ame ri-
can Indians. In D . M . Schullian and M. Schoen, edito rs, Music and medicine ,
pp . 25 - 46. New York: Hen ry Schuman.
Depont, Octave, and Zavier Coppolani. 1887. Les confreries religieuses musu/ -
manes . Algie r: Adolphe Jou rd an.
Dermenghem, Emile, and Leo-Louis Barbes. 195 1. Essai sur la hadhr a des Ais-
saoua d'Algerie , R evue africaine 95 :289 - 314.
f Devereux, Georges. 1970. Ess ai d'ethnopsychiatrie generate. Tr anslated by Tin a
Jol as and Henri Gobard. Paris: Gallimard.
___ . 1983 . La crise initi atique du chaman chez Platon (Phedre 244 d- e). Psy -
chiatriefranc;aise, n.s. 6:33 -3 5 .
Dhorme, Edouard. 1956. Intr oduction a La Bible. Bibliorheque de la Pleiade. Par-
is: N.R. F., Gallimard.
Diererlen, Germaine. 1941. Les dmes des Dogons. Paris: lnstitut d'Ethnologie.
___ . 1951. Essai sur la religion bambara. Paris: P.U .F.
Dodds , E. R. 1951 . Th e Greeks and the irrational. Berkeley : University of Cali-
fornia Press .
___ . 1960. Eurip ides, Baccha e. Edited with introduction and commentary by
E . R. Dodds. Second edition. O xfo rd: T he Clarendon Press.
Dreyfus -Roche, Simone. 1954 . Chants indiens du Venezuela. Seance de chaman -
isme. Cf. discog raph y.
Dumont, Louis. 19 57. Une sous -caste de l'lnd e du Sud. Paris-La H aye: Mouton.
During , Jean. 197 5. Element s spiriruels clans la mu sique tradi tionn elle contem-
poraine . Sophia Perennis 1, no . 2:129 - 54. Teheran: Imperial Iranian Academ y
of Philosoph y.
Eliade, Mircea . 1948. Techniques du Yoga . Paris : Gallima rd.
___ . 1951. Le chamanisme et /es techniques archaiques de /'extase, 2nd ed .
1969. Paris: Payor.
Bibliography

Emm anuel , Maurice. 1914. Gr ece (art greco-romain ). In Enc yclo pedie de la musi-
359 l
que et Dictionna ire du Con ser vat oi re l :37 7- 537. Paris: Delagrave.
England, icholas M. 196 7. Bushman counte rp oint. Journal of the Inte rnational
Folk Music Coun cil 19:5 8- 66.
Erlanger , Rodolphe d' . 193 0. La m usique arabe, vol. 1. Al-Farabi , Kitab al-
musiqi al-kabir (Grand tr aitc: de la musique ), books 1 an d 2 . Paris: Paul
Geuthner.
___ . 1935. La musique arabe, vo l. 2. AJ-Farabi, book 3 and Avicenna , Kitab
al-shifo' (Livre de la gue riso n ) (mathematiques, ch ap. 12 ). Paris: Paul Geuthner.
___ . 193 8 . La musique arabe, vo l. 3 . ~afI al-D in, 1: A l-sharafiyya (Ep icre a
Sha raf al-Di n ); 2 Kitab al-adwar (Livre de s cycles music aux ) (Comm entaires ).
Paris: Paul Geuthner.
E uripides . Bacchae; H eracles; H ippolytus .
Evans -Pritchard, E . E. 1937 . Witchcraft, oracles and magic among the Azande.
Oxford: Cla rendon Press.
Farabi, Al-. See Erl anger 1930 and 1935.
Farmer, H enry George . 19 08 . Arc. (Musik i). In Enc yclopedie de /'Islam.
___ . 1926 . The influence of music : From Arabic sources : A lecture delivered
before the Musical Association. London : H arold Reeves.
___ . 1929. A history of Arabian music to the X III th century. London: Luz ac
and Co .
___ . 1940 . Th e sources of Arabian music : An annotated bibliography . . . ,
Glasgow.
___ . 1957. The music of an cient Mesopotamia. In The new Oxfo rd history of
music l :228 - 54. London: Oxford Unive rsit y Press.
___ . 1959. Th e science of music in the Mafafih al-'U lum. Transactions of the
Glasgow University Oriental Society 17, years 1957 - 58 .
Felice , Philippe de. 1936 . Poisons sacres . lvresses divines . Essai sur qu elq ues
formes inferieures de la m ys tique . Paris: Albin Michel.
___ . 194 7 . Foules en de/ire , extases collectives. Essai sur quelques fo rm es in-
ferieures de la m yst ique. Paris: Albin Miche l.
Fer chiou, Sophie . 1972. Sur vivances mystiques et culte de possession clans le mar-
abourisIT'e cunisien. L'Homm !J: R evuefram;aise d'anthropologie 12 :47- 69 .
Field, Margaret J. 196 9. Spirit possession in Gh ana . In John Beattie and John
Middleton, editors , Spirit mediumship and society in Africa, pp. 3-13. Lon -
don: Routledge and Kegan Paul.
Filiozat, Jean. 196 3 . La n ature du yoga clans sa tr ad ition. In T herese Bres se,
Etud es instrumentales des techniques du yoga, pp. i- xxvi ii.
Firth, Ra ymond. 1969. Foreword to John Beattie and J ohn Middleton, eds ., Spirit
mediumship and society in Africa . London : Routle dge and Kegan Paul.
Foucault, Michel. 1961. Folie et deraison. Histoire de la Jolie a l'dge classique .
Paris : Plon .
Frances , Robert. 1958. La perception de la musique. 2n d ed. 1972. Paris : Li-
brairie philosophique Vrin .
Frisbie, Charlotte J. 1971 . Anthropological and erhnomusicologic al implications
of a comparat ive analysis of Bushman and African Pygmy music. Ethnolog y
10 :265 - 90 .
 360 BIBLIOGRAPHY

Gab o rie a u, Marc. 1975. La tr a nse ricuelle clans !'H imala ya cent r al: fo lie, ava t a r ,
medication. Purasartha. Recherches de sciences sociales sur l'Asie du Sud (pan
2 ), pp . 147 - 72 . Pari s : Centre d'Ecudes de l'lnde er de l'Asie du Sud.
Gallini, Cla ra . 1967. I rituali dell'argia . Pad u a : C.E .D .A.M. (with an append ix,
"L a musica nei ricuali dell 'arg ia" by Die go Carpirella and a recording 17 / 33 ).
Ganav, Solange de. 1942. Le Binou yebene (Miscellanea Africana Lebaud y, cahier
no . 2 ). Pari s : Libr. o r ient. Paul Geuchner.
Garder, Louis. 1952 a nd 1953. Un probleme de mystique compa ree: la m enti on
du nom di,·in (dhikr ) clans la m ystique musulmane. Revue thomiste 52, no.
1 :642 - 79 and 53, no. 1 :197 - 2 16 .
___ . 1976 . See Anawari .
Gessain, Robert, and P. E. V icror. 1973 . Le cambour chez les Amm assali m iur
(ccne est du Groen land) . Objets et Mon des 13 : 129-60. ·
r Ghazzali, Abu H amid Al -. See MacDonald.
Ghazzal1, A~mad al - (Majd al -Din ). See Robson.
Gra y, Robert F. 1969. T he Shecani cult among rh e Segeju of T anzan ia . In John
Beanie and John Middleron, ediro rs, Spirit mediumship and society in Africa,
pp. 171 - 187 . London : Rourledge and Kegan Paul.
Grebe, Maria E ster. 1973. El Kultrun m apuche : un microcosmo simbolico. Re -
vista Musical Chilena 2 7 , nos. 123-24:3-42.
H ama,·on, Roberre . 1978a . Soigner le m ore pour guerir le vif . Nouvelle revue de
psychanalyse 17 :55- 72 .
___ . 1978b . Les heros de serv ice. L'Ho111me: R evuefranc;a ise d'anthropologie
18:17-45.
H assan, Scheherazade Q assi m. 1975. L es instruments de musique en lrak et leur
role dans la societe traditionnelle (roneor ype ). Paris : E.H.E.S.S. Now pub -
lished under rhe same title by Mouton Edireur , Pa ris 1980 .
__ _ . 1977. Les instruments de m us iqu e chez !es Yezidi de l'lrak. 1976 Year-
book of the International Folk 1\tIusic Council, pp. 53-72.
H enderson, Isobel. 1957. Ancient Greek music. In ,'\·ew Oxford history of music
1:336 - 403.
Henney, Jeannette H. 1973. The Shake rs of Sr. Vincent : A stable reli gion . In E.
Bourguignon, ediro r, Religion, altered states of consciousne ss and social
change, pp . 219 - 69 .
H erodotus . 1945 . Hist oires . T exr estab lished and rr anslared by P. E. Le gran d .
Paris: Soc iere d'edirion "Les Belles Lemes ."
H ersko,·irs, Melville J. 1938. Dahome): an ancient \Vest African kingdom. 2 vols .
E ,·ansron, Norrh,,·esrern Uniw rsiry Pre ss .
___ . 1943a . Pesquisas etnologicas na Bahia. Bahia : Museu d'Esrado .
___ . 1943b . The souchernmosr outpost of Ne\1· World african ism . American
A nthropologist 45 :4 95 - 510.
Herrz, Robert. 192 . Secres russes (C. R . de K. K . Grass . Die russischen Sekten ) .
.\I elanges de sociologie rel1gieuse et folklore. Paris : Felix Alcan.
Heusch, Luc de. 1971 . Possession er chamanisme (1964 ) and La folie des <lieux er
la raison des hommes. In Pol!rquoi /'epouser? et autres essa is. Paris: Gallima rd .
Hitchc ock, J ohn T. , a nd Rex L. Jones . 1976 . Spirit possession in the Nepal H ima -
layas . Warminster (England ): Aris and Phillips.
Bibliog raphy
36 1
Holt ved, Erik . 1967 . E sk im o sha manism . In C. M. Edsman, editor, Studies in
shamanism. Stockholm .
Horton, Robin. 1969 . T yp es of spirit possession in Kalabari religion . In John
Beattie and John Middleron, editors, Spirit mediumsh ip and society in Africa,
pp. 14-49 . London : Routledge and Kega n Paul.
Hu sso n, Raoul. 1960. La vo ix cha nt ee. Paris: Gaurh ier-Vill ars.
Jackson, Anth ony . 1968. Sound and ritu al. Ma n, n. s., 3:293 - 99.
Jacob, Franc;ois. 1970 . La log iqu e du vivant . Une histoire de l'heredite . Paris : Gal -
limard.
Jakobson, Rom a n. 1960. Linguistics and poetics . In T . A. Sebeok, ed itor, Style in
language . Camb ridge, Mass .: M .I.T. Press.
]arg y, Simon . 1971 . La musique arabe (coll. "Que sais -je ' " ). Pa ris: P.U .F.
Je a nmaire, H. 1939. Couroi et Cour ete s. Es sai sur /'education spartiate et sur
/es rites d'adolescence dans /'antiquite hellenique . Avesnes -su r-H elpe: G. Del-
off re.
_ __ . 1949a. Le tr aitemenr de la mania clans les "Mysteres " de Dionysos et des
Corybanres. j ournal de Psych olog ie 64 - 82.
__ _ . 1949b . Le sa tyre et la menade . Remarques sur quelques cextes rel atifs aux
d anses "orgiaq ues. " In Melang es d'archeologie et d 'histoire of(erts a Charles
Picard . . 1 :463 - 73 . Paris: P.U .F.
_ __ . 1951. Dionysos . Hist oire du culte de Bacch11s. Par is: Payo r.
Je rvis, Giovanni. 1966 . Considerations neuro -psychiatriques su r le tarenrulisme .
In E . De Martino, La terre du rem o rds, pp . 329 - 54. Paris: Gallima rd.
Jette, Fern and. 1961. L'exrase clans la tr ad ition spiritue lle du xiiie au xv iie siecle .
Di ctionnaire de spiritualit e, pp . 21 3 1-4 7 . Paris: Beauc hesne.
John ston, Th omas F. 1972 . Possession music of the Shangana -Tsonga . Af rican
.\f 11sic, 5 :1 0- 22.
Jo rd ania, Red jeb . 1978 . Boris V ian et la chan so n. Musiqu e en jeu 30 :91 - 110 .
Juillerar. Bern ard. 1977 . Introduction au numero : "Folie," po ssess io n et chaman -
ism e en Nom·elle Guinee. Jou rna l de la Societe des Oc eani st es 33 :117 - 21.
Junod, Henri A. 1913. T he life of a South Afr ican tribe. 2 vo ls . Neufchatel: lmpri-
merie Attinger Freres .
Jun od, H enr,· P. 1934. Les cas de po sessio n et l'e xorcisme che z les Vat dau. Afri -
Co1 7 :27 0- 99 .
Kakouri . Kaieri na J . 1965 . Di onysiaka . As pects of the Thracian religion of to -
da,'. Athen s: G . C. Eleftheroud akis .
Karcomi, i\1argaret. 1973 . .viusic and tr ance in centr al Ja,·a. Et h11omusicology
17 :16 3- 208 .
Kaiib, al-H asan .-\.1 -. See Shiloah .
Kilson . .\J arion. 1971 . Kpele Lala . Ga religions, songs and symbols. Ca mbridge ,
Mass .: H a n ·ard Univers in · Press .
Kindi .. -\.I-. See Co\\' l.
Kirchma,·er, Jean. 1961. Ex case chez Jes Peres de l'Eglise. Dictinnnaire de spiritua -
lite, s .,·. "Excase .'' Pari s: Beauchesne .
Kiri , R . P. Gab riel. 1968 . Consecration a un fetiche au Daho meY. Etudes Dah o-
meennes, 11 :26 - 32 .
Knox, R. A . 195 0 . Enthusi asm, a chapter in the history of religion. Oxfo rd.
I,

362 BIBLIOGRAPHY

Kurath, Ger tru de. 1960 . Panora ma of d ance ethnology . Current Anthr opo logy
1 :233 - 54 .
Laca rri ere , Jac ques. 1955. L'epouvante : Th eat re et possession au temps d' E s-
ch yle. In E schyle et /'O restie (Ca hiers de la Compagnie Madeleine Renaud -Jean -
l ouis Barrault ), 3e annee, 1 1e cahie r, 58 - 65 . Pari s: Jull iard .
___ . 1961. Les hommes ivres de Dieu. Pari s: Arrhaud .
Laloum, Claude, and Gil be rt Rouge t. 1965. D eux chants liturg iques yo ruba.
journal de la Societe des Africanistes 35 : 108 - 38.
Lapassade, Georges . 197 5. T ranse et possession au Maroc. Que/ corps? 2: 27 - 30.
_ __ . 1976 . Essai sur la transe. Paris : Je an-Pierre D elarge .
Lapl anche, J., and J . B. Pontalis. 1968 . Vocabula ire de la psychanaly se. Pari s:
P.U.F.
Lasserre, Franc;:ois. 1954. Plutarch, De la musique . T exte, tr ad uction, co mm en -
taire precede d'une etude sur !'education musicale clans la Grece anti que (Biblio -
theca H elvetica Romana ). Olten and Lausanne: Urs G raf-Verlag .
La\\·ler, Lilian B. 1927 . The m aenads : A contribution to the study of dance in an -
cient Greece . 1vl emoirs of the American Academy in Rome , 6 :69 - 11 2 .
Lee . Richard B. 1968 . The sociologv of !Ku ng Bushman trance performances . In
R . Prince, editor, Trance and possession states, pp. 35 - 54. Montreal : R. M .
Bucke ~vl emorial Societv .
Lee , S. G . 1969 . Spirit possession among the Zulu . In John Beattie and John
Middleton, editors . Spirit mediwnsh ip and society 111 Africa, pp . 128 - 56. Lon -
don : Routledge and Kegan Paul.
Le Herisse, A . 1911 . L'ancien royaume du Dahomey. Pari s: Emile Larose .
Leiris , Michel. 1934 . L'Afriquefantome. Paris: Gallimard .
_ __ . 1955. Fourbis. Paris : Gallimard .
___ . 1958 . La po ssessi on et ses aspects theatraux chez /es Ethiopiens de Gon -
dar. Paris : Pion .
___ . 1974. Mazmur le clerc . L'Ethnogr aphie , n .s., 68 :39 - 58.
Le Jeune, Claude. 1598. D udecacorde contenant douze psaume s de David .
Reproduced in H enry E xpert, Les maitres musiciens de la Ren aissance fran -
<;aise. Par is : Leduc . 190 0 .
Lesure, Franc;:ois. 1961 . Claude Le Jeune. In Encyclopedie de la musique l/l. Paris:
Fasquelle.
___ . 1966 . _\Jusica e societa . Milan : lsrituto editoriale italiano.
Levi -Strauss, Claude . 1949 . L'efficacire symbolique . Anthropologie structura le.
Pion: Paris, 1958.
___ a
. 1960 . Introduction !'oeuvre de Marcel Mauss . In Marcel Mauss, Socio -
logie et Ant hropologie . Pa ris: P. U .F.
___ . 196 2. La pensee sauvage. Paris: Pion.
___ . 1966 . Du miel aux cendres (Mythologiques II). Paris : Pio n.
Le\,-is , I. M _ 1971 . E cstatic religion: An anthropological study of spirit possession
and shamanism. H ar m ondswo rth: Penguin Books .
Linforth, Ivan M . 1946a . The coryb antic rites in Plato . University of California
Publications in Classical Philology. 13, no. 5, 121-62. Berkeley and Los Ange -
les : Universit\· of California Press.
___ . 1946b . Telestic madnes s in Plato , Phaedrus 244 DE. Un iversity of Cali -
fornia Publications in Classical Philolog)\ 13 , no . 6, 163 - 72 .
Bibliograph y 363
Lomax , Alan. 1968. Folk song style and culture. Wash ington : American Associ-
ation for t he Advancement of Sciences.
Lor -Falck, Eveline. 1961 . L'anim ation du tambour. Journal asiatique yea r
1961:213 - 39.
___ . 1968. La divination clans l'Arctique et l'Asie septentrionale. In Caquot
and Leibovici, editors, La divination 2:247-77. Paris: P.U .F.
___ . 1970. Psychoparhes er chamans yakou tes . In £changes et communi ca-
tions. Melanges offerts a Claude Levi-Strauss a /'occasion de son 60 e anniver -
saire, reunis par J. Pouillon er P. Maranda. 1: 115-29. The H ague -Pa ris: Mou-
ton.
___ . 1977. Apropos du terme chamane. Etud es mango/es . .. et siberiennes ,
cahie r 8, 7-1 8. Nanterre, Labethno: Unive rsite de Paris X.
MacDonald, Charles. 1973. De quelques manifestations chamanistiques a
Pala wan . In Asiedu Sud -Est et monde insulindien, Bulletin du Centre de docu -
mentation et de recherche (CeDRASEMI ) "Chamanisme er possession" (fasci -
cule 2 ), vol. 4, no. 3:11 - 18. Paris : Mouton.
MacDonald, Duncan B. 1901 and 1902. Emotional religi on in Islam as affected by
music and singing . Being a translation of a book of the Ihya 'UIZ:imad -Uin of
al-Ghazzali with analysis , annotation, and appendices . Journal of the Royal
Asiatic Society 1901, 195- 252 and 705 - 48; 1902, 1-28.
___ . 1908. Sama. In Encyclopedie de /'Islam.
___ . 1908 . Dhikr. In Enc yclopedie de /'Islam .
Majd al -Din, Ahmad al-T usi Al-Ghazzali. See Robson.
Mars. Louis. 1953. Nouvelle contribution a l'erude de la crise de possession. In
Les Afro -ame'ricains (Memoires de l'Instirut Franc ais d'Afrique Noire, no. 27 ),
pp. 213 - 33. D aka r : I.F.A. .
Marshall, Lorna . 1969. The medicine dance of the !Kung Bushmen. Africa
39:347 - 81.
Masaraki, Despin a W. 1974. Ein Aulos der Sammlung Karapanos. Mitteilungen
des Deutschen Archaologischen Institut Athenische Abteilung 89: 105 - 21.
Massignon, Louis. 1922 . Essai sur /es origines du lexique technique de la mystique
musulmane. Paris: Paul Geuthner.
___ . 1934. T arTka. In Enc yclopedie de /'Islam.
Mauss, Marcel. 1936. Les techniques du corps. Reprinted in Sociologie et anthro -
pologie, 1960. Paris: P.U.F .
Meinecke, Bruno. 1948. Music and medicine in classical antiquity. In D. M.
Schullian and M. Schoen, editors , Music and medicine, pp. 47 - 97. New York :
Henry Schuman.
Merriam, Alan P. 1964. The anthropology of music. Evanston: orthw estern
Univ er sity Press .
Mersenne, Marin. 1636. Harmonie Universe/le. Edited by Fran~ois Lesure, 3
vols. Paris : C. .R.S.
___ . 1932. Correspondance. Edited by P. Tanner y.
Merraux, Alfred . 1942. Le chamanisme araucan. In Religions et magies indiennes
d'Ame'rique du Sud, ed . Simone Dre yfus, Paris : Gallimard, 1967.
___ . 1958. Le va udou hai'tien . Paris: Gallimard.
Michelet, Ju les . 1862 . La sorciere .
364 BIBLIOGRAPHY

Misson, Maximilien . 1707. L e theatre sacre des Cevennes. London: Robert Rog -
er. Reissued by Les Presses du Languedoc. Brignon: 1978.
Mokri, Mohammad. 1961 . Soufisme (Le S. er la musique ). In En cyclopedie de la
musique 3:1014-15. Paris, Fasquelle ..
M ole, Marijan . 1963. La danse exrarique en Isla m. In Les danses sacrees (coll.
"Sou rces orienrales" 6), pp. 145 - 280 . Paris: Editions du Seuil.
Molino, Jean . 1975 . Fa ir musical er semiologie de la musique. Musique en Jeu
17 :37-61 .
.tvlonfouga -Nicolas, J acquel ine. 1972 . Ambivalence et mite de possession. Paris:
Editions Anrhropos .
Morechand, Guy. 1968. Le chamanisme des Hm ong . Bulletin de /'Ecole Fram,a ise
d'Extreme -Orient 54:53 -2 94 .
Moursopoulos, E\·anghelos . 1959 . La M usique dans /'oeuvre de Piaton . Par is:
P.U.F.
Nadel, S. F. 1946. A study of shamanism in rhe Nuba Mountains. Journal of the
Royal Anthropological Institute of Great Britain and Ireland 76, 1, 25-37 .
. arriez, Jean -Jacques . 197 5 . Fondements d'une semiologi e de la musique (coll.
'· 10/18' ' ). Paris: Union generale d'edirion .
Nebesky-Wojkowirz, Rene de . 1956 . Oracl es and demons of Ti bet . T he H ague :
Mouton.
Needham, Rodney. 1967. Percussion and transition . i',fan, n .s., 2 :606 -14.
eher, Andrew. 1961. Auditory driving obse rved wirh scalp electrodes in normal
subject. El ectroenc ephalography and Clinical >ieurophysiology 13 :449-51 .
___ . 1962. A physiological explanation of unusual behavior in ceremonies in-
rnlving drums . Hum an biology 4:151 - 60.
1 errl, Bruno . 1977. A propos des universaux . Le Monde de la Musique / Th e
\'Corld of i\fusic 19: - 12 .
Nkeria, J. H. 1957. Possession dances in Afric an societies . journal of the Int erna -
tional Folk Music Council 9:4- 9.
Nostradamus, Cesar. 1614 . Histoire et chronique de Provence. . Lyo n : Rigaud .
Oesterreich, T. K. 193 0 . Possession, demoniacal and other, among primitive
races, in antiquit)\ the middle -age and modern tim es (autho rized translation of
rhe 1921 original in German ).
Oppirz, lvlich ael. 19 1. Shaman en in Blind en Land. Frankfurt -a m-M a in: Syndi -
kar.
Pages, ~lax. 1973. La libe ration du corps. L'h o111111e et la societe. Revu e interna -
tionale de recherches et de syntheses sociologiques 29 - 30 : 15 3- 7 5 .
Pairaulr, Claude . 1966. Bo11111 -le-Grand, village d'lro. Paris: ln srirur d'Erhnolo -
g1e.
Pearson .. \. C. 1918. Possession (Greek and Roman ). In Encyclopaedia of Reli -
gions ,111dEthics l 0: 127-30 . Edinburgh: James Hastings.
Pidoux, Dr. Charles. 1955. Les erars de possession riruelle chez !es Melano -Afri -
cains . L'evol11tio11psychiatrique l l: 271- 3.
Plato. Crito; Ion; Laches ; Laws; Phaedrus; Philebus; Republic ; Symposium;
Th eaetetus; Tima eus .
Plutarch . 1953. Dialogu e sur /'amour (Eroticos ). Text and tran slation 'by Robert
Flacelie re. Paris: Sociere d'edirion "Les Belles Lerrres ..,
Bibliography 365
Parser, Charles. 1970 . See Rousseau 1781 .
Portier , Richard. 1973. ores sur !es chamanes er mediums de quelq u es groupes
Th ai .Asi edu Sud -Est et Monde Insulindien. Bulletin du CeDRAS EMI 4, no. 1.
Paris : C.N.R .S. er E.P.H.E.
Pouillo n , Je an . 1972. Ma lade er medecin : le meme er / ou l'aurre? (remarques erh -
nologiques ). Nouvelle Rez;ue de Psycha nalyse 1: 77- 98.
Prince, Raymond, ed itor. 1968. Tranc e and possessicm states (Proceedings of rhe
Second Annual Conference, R. M. Bucke Memorial Society 1966 ). Monr r eal.
Prince, Raymond. 1968. Can rhe EGG be usoo in rhe srudy of possession scares? In
R. Prince, editor, Trance and possession states, pp. 121-37.
Pseudo -Plut a rch. See Lasserre.
Quenum, Maximilien . 1938. Au pays des Fons, us et coutumes du Dahomey. Pa r-
is: Emile Larose.
Rausky, Franklin. 1977 . Mes m er ou La rez;olution therapeutique. Paris: Pa yo r.
Reinach, Th eodo re. 1900 . La musique des spheres . Re vue des etudes grecques
13 :432 -4 9 .
Reinhard, Johan. 1976. Shamanism an d spirit possession : Th e definition prob-
lem. In John T. Hitchcock and Rex L. Jones, editors, Spirit possession in the Ne -
pal Himal ayas, pp . 12-20. \Xiarminsrer : Aris and Phillips.
Robson, James. 1938. Tracts on listening to music. Being Dhamm al-malahi" by
lbn ab 1<I-Dun ya and Bawa nq al-ilma c by \1 ajd al -Din al-Tusi a l-Ghazzal, (coll.
·'Oriemal T ranslarion Fund," n.s. , vol. 34 ). London: The Raval :\si ar ic Society.
Rodrigues, Nina. 1935. 0 animismo fe tichista dos negros bahianos (Prefacio e
noras de Arthur Ramos ). Civilisa~ao Brasileira, Rio de Janeiro. (First published,
in insrallmems, in Rez;ista Brasileira , in 1896-97. T ranslared and published in
French in 1900 ).
Roger, Joseph Louis. 1803. Tra ite des effets de la musique sur le corp s hu -
main . . . Translated from rhe Larin edition (17 4 8) by Etienne Sainre -Marie.
Ronsard, P. de. 1866. Oeuvres completes . Published by M. Prosper Bla nchemain.
Paris: Librairie A. Fr anck, 1866.
Rouaner, Jules . 1922. La musique arabe. In Encyclopedie de la musique et dic -
tionnaire du Conservatoire 5: 2676 - 2939. Paris: Delagrave.
Rouc h, Jean. 1960. La religion et la magie Song hay. Paris: P. U . F.
_ __ . 1973. Essai sur !es avarars de la personne du possede, du magicien, du
sorcie r, du cine as re er de l'erhnographe. In La notion de personne en Afrique
noire. Collogues lnrernarionaux du C.N.R .S., no. 544:529 - 44 . Pa ris: E ditions
du C.N.R.S.
Rouger, Gilbert . 1957. Inrroducrion ro ates sur la musique des Bochiman com -
paree a celle des Pygmees Babinga, erablies par Yverre Grimaud avec la collabo -
ration de G. Rouger . Paris: Musee de !'H omme, Departemenr d'erhno -musico -
logie and Cambridge: Harvard Un i\·ersiry, Peabod~· Museum . Nares ro rhe
record; cf . discograph\ ·.
_ _ _ . 1961 . Un chromarisme africain . L'Homme, Revu e f ranc,aise d'anth ropo -
logie l :30 - 46 .
___ . 1962. Musique '·vod un " (Dahomey ). Actes du VI e Congres International
d es Sciences Anthropologiqu es et Ethn ologiques (Paris 1960 ) 2, 2: 121 - 22 . Par -
is.
 366 BIBLIOGRAPHY
___ . 1965. Notes et documents pour servir a l'etude de la rnusi que yorub a.
Journal de la Societe des Afr icanistes 35 :67 - 107.
___ . 1971. Court songs and traditional history in the ancient kingdoms of ·
Porto-Novo and Aborne y. In K. P. Wachsrnann, editor, Essays on Music and
Hi story in Africa, pp. 27- 64, Evanston: Northwestern University Press.
Rousseau, Jean-Jacques. 1765. Musique. In Diderot , editor, Enc yclopedie ou Dic-
tionnaire raisonne des sciences, des arts et des metiers, vol. 10.
_ __ . 1768. Dictionnaire de la musique. Paris .
.,
\
___ . 178 1. Essai sur l'origine des langues, ou ii est parli de la melodie et de
/'imitation musicale. Edited by C. Porset. Bordeaux: Ducros, 1970.
!'I·
t. ___ . 1782. Supplement a l'Encyclopedie, vol. 3. Neufchatel and Amsterda m .
il~ _
ii,!
Roux, Alain. 19 73. La rnusique pop. In Paul Beaud and Alfred Willener, Musique
et vie quotidienne. Essai de sociologie d'une nouvelle culture. Paris: Marne.
Roux, Jean-Paul. 1963. La danse charnanique de l'Asie centrale. In Les danses sa-
,,
'I' crees, pp. 281 - 314. Paris: Editions du SeuJ .
Sachs, Cun. 1940. The history of musical instruments. New York: W. W. Nor-
ton.
$afi al-Din. See Erl ang er 1938.
Schaeffner, Andre. 19 51. Les Kissi. Une societe noire et ses instruments de musi-
que. Paris : Hermann.
___ . 1965. Rimel et pre-theatre . Hi stoire des Spectacles (Enc yclopedie de la
Pleiade ), pp . 21-54. Paris: Gallirnard.
Schneider, Marius. 1948. La danza de espadas y la tarentela. Ensa yo musicol6-
gico, etnografico y arqueologico sabre los ritos medicinales. Barcelona: Insti-
tute Espanol de Musicologia.
Schott-Billrnann, France. 1977. Corps et possession. Le vecu corporel des pos -
sedes face a la rationalite occidentale . Paris: Gauthier -Villa rs .
Schullian, D. M. and M . Schoen, editors. 1948 . Music and medicine. New York:
Henr y Schuman.
Sechan, Louis . 1930 . La danse grecque antique. Paris: E. de Boccard .
Segurola , R. P. B. 1963. Dicti onnai re fon-franc;ais (2 vols ., roneotype ). Cotonou:
Procure de l'Archidiocese.
Sergent, Alain. 1952. Leroi des prestidigitateurs, Robert-Houdin. Paris: Editions
du Seuil.
Seydou, Christiane. 1972. Silamaka et Poullori. Re cit ipiq ue peul raconti par Tin -
guidji (Classiques africains ). Paris: A. Colin.
Shiloah, Arnnon. 1965 and 1967. L'Ep itre sur la rnusique des Ikh wa n-al-Safa. Re-
vue des Etudes Islamiques 32:125 - 62; 34:159 - 93.
--- · 1972. Al-r:fasan ibn Afirnad ibn 'Ali al-Kat ib. La perfection des connais -
sances musicales (Kitab Kam al adab al-ghina ). Tr aduc tion et cornrnentaire d'un
traite de rnusique arabe du xie siecle. Paris: Paul Geuthner.
Shirokogoroff , S. M. 1935. Psychomental complex of the Tungus. London: Ke-
gan Paul.
Sigerist, Henr y E. 1948. The stor y of tarencisrn. In D. M. Schullian and M.
Schoen, editors, Music and medicine, pp. 96-11 6. New York: Henry Schum an.
Simon, Gerard. 1979 . Kepler astonome astrologue . Paris: Gallirnard.
Bibliography 367
Simon , Pierre, and Ida Simon-Barouh. 19 73. HauBong . Un culte vietnamien de
possession transplante en France. Paris: Mouton.
Stein, R. A. 1962 . La civilisation tibetaine . Paris: Dunod.
Srurre va nt , W . C. 1968. Categories , percussion and ph ysiology. Man , n.s.,
3:133 - 34.
Surugue, Bernard. 1972. Contribution a l'etude de la musique sacree Zarma
Songha y (R epubl ique du Niger ). Erud es Nigeriennes 30, Niamey.
T anner , R. E. S. 1955. H ysteri a in Sukuma medical practice. Africa 25:2 74- 79 .
T ardies, Claude. 1958. Porto-No vo . Les nouvelles generations af rrcaines entre
leurs traditions et !'occident. Par is, La H aye: Moucon.
T heresa of Avila (Sainte T herese de Jesus ). 1949. Oeuvres completes. Transl ated
by R. P. G regoire de Saint Joseph. Paris: Edition s du Seuil. Obras comp/etas .
1948. Madrid: Editori al Plenirud.
Th omas, Anus, Seig. d'Ambry . 1603. Ode su r la musique du defunct sieur Clau-
dio Le Jeune. Reproduced in Henr y Exp ert, L es maitres musiciens de la R enais -
sance fra nqaise (Claude Le Jeune, Le Printemps, Jer fascicule ). Paris: Leduc,
1900 .
___ . 1611. "Commentaires" figurant clans la craduccion , pa r B. de Vigenere,
de : Philoscrace, De la vie d'Apollonius de T hyaneen. Paris: chez la veuve M acch-
ieu Guillemoc .. ..
Tidjani, Serpos. 1950. Riruels. In Le monde noir, Presence afncaine, numero spe -
special 8-9 :29 7- 305 . Paris .
Torgue, Henry Skoff. 197 5. La pop -music (Collection "Que sais - je l" ). Par is :
P.U .F.
Touma, H abib H assan . 1976. Quelques noces sur Jes rapports ent re eschecique er
improvisation clans la musique arabe . The \Y/orld of Musi c. 18:51 - 52 .
TraceY, Andrew. 1970. T he Macepe Mbira music of Rhodesia . African Music
4:37 - 61.
Trimingh am, J. Spencer. 1949. Islam in the Sudan. London : Oxford Unive rsity
Press.
___ . 1971. Th e Sufi orders in Islam. London: Oxford Unive rsit y Press .
Turnbull, Colin M. 1961. Forest people. New York: Simon and Schuster.
Turner, Viccor W. 1968. The drnms of affliction. A study of religious process
among the 1':dembu of Zam bia. Oxford : Clarendon Press.
Tyard, Pontus de . 1555. Solitaire second, ou Prose de la .Musique. Lyons : Je an de
Tournes.
Vasib·ic, G. :VL 1969. E venki . Istoriko-etnograficeskie oce rki. (XVIII -nacalo
XX\· .). (Les Evenki . ..\perc;us hisrorico -echnog raphiques du xviiie au debut du
xxe s. ). Leningrad: Izd. Nauka.
Veith, Ilza. 1973 . Histoir e de /'hysterie . Tr ans!aced by Sylvie Dreyfus. Paris :
Seghers .
Verger, Pierre . 1954. Ro le joue par l'erac d'heberude au cours de !'initiation des
novices aux Cultes des Orisha er des Vodun. Bulletin de l'Institut Franqais d 'A -
f rique _ oire 16 :322 - 40 .
___ . 1957 . Notes sur le culte des orisa et vodun (Memoires de l'lnscicuc Fr an-
<_;ais
d'A.frique Noire, no . 51 ). Dakar: I.F.A.N.
 368 BIBLIOGRAPH

__ _ . 1969. T ranee and convention in Nago-Yoruba spirit mediumship. In Joi-

Beattie and John Middleton, editors, Spirit mediumship and society in Afric .
pp . 50 - 66. London: Routledge and Kegan Paul.
Vernanc, J. P. 1974. Paroles et signes muets. In J. P. Vernant, editor, Divination
rationalite. Paris : Editions du Seuil.
___ . 197 5. Image et apparence clans la thforie platonicienne de la mimesz
Journal de psychologie 2:133 - 60.
Vian, Boris. 1958. En avant la zizique ... et par ici /es gros sous . Paris: Le Liv
concemporain.
Vincent, Jeanne·Fran<;0ise . 1971. Divination et possession chez !es Mofu, mo:
l- tagnards du Nord-Cameroun. Journal de la Societe des Africanistes 41: 71 -1 3:
r. Vitray -Me yerovitch, Eva de. 1972. Mystique et poesie en Islam. Djalal-ud -D
..'-".'
t.;i"
,
I- ·1
Rumz et /'Ord re des Der viches tourneurs. Desclee de Brouwer.
Wachsmann, Klaus P. 1957. See Nketia, J. H.
j: ___ . 1971. Universal perspectives in music. Ethnomusicology 15 :3 81-84.
1!
Walker, Sheila S. 1972. Ceremonial spirit possession in Af rica and Afro-Americ
Leiden: E. J. Brill.
Wilamowitz, U. von. 1920. Plato. 2d. ed. Berlin.
Yasser, J. 1926. Musical moments in the shamanistic rites of the Siberian pag,
tribes. Pro-Musica Quart erly4 :4-15.
Yates, Frances A. 194 7 . T he French academies of the sixteenth centu ry (Studies
the Warburg Institute , vol. 15 ). London.
Zemp , Hugo. 1978. 'Are'are classification of musical types and instruments. Et
nomusicology 22:37 - 67.
Zempleni, Andras. 1966. La dimension therapeurique du culte des rab, ndi:i
tuuru et samp; rites de possession chez les Lebou et les Wolof. Psychopa tholo[;
africaine 2:295-439.
__ . 1972. See Adler.

! ,.