~ ~C

i~

t'""

>::>

~>

~ tTl

~~

ZC)

o>

~o

-~

~Z

~t!

j"'t

l('b

~~

o..~

ZtT

l

>CJ

)C)

~CJ

)('b

~.-

~n

Z-

~o ~

~":rj

~~

Uo ~ o

~ ~ ~

I CONTENTS

List of Figures xx

I I. The Problem Defined:The Need for an Anthropologyof Art I

II. I. Can there be an Anthropological Theory of Visual Art? I

1.2. The Art Object 51.3. Art Sociology 71.4. The SilhoUdtt: of an :\nthropologlc;;.l Thcory 10

I 2. Thc Theory ofthc Art Nexus 122. I. Construeting a Theory: Terms and Relations 122.2. The Index 13

I 2.3. Abduction 14

I 2.4. The Social Agent 162.5. 'Things' as Social Agents 17

I 2.5.1. Paradox Elimination 192.5.2. Agents and Patients 21

I 2.6. The Artist 232.7. The Recipient 242.8. The Prototype 252.1). Summary 26

3. The Art Nexus anil the Index 283. I. The Table of AgentJPatient Relations between Four Basic Terms 283.2. Index-A ~ Artist-P 283.3. Index-A ~ Recipient-P 313.4. Index-A ~ Prototype-P 323.5. Artist-A ~ Index-P 333.6. Recipient-A ~ Index-P 333.7. Prototype-A ~ Index-P 353.8. The Centrality ofthe Index 35

3.8. I. The Logic of 'Primary' and 'Secondary' Agentsand Patients 36

3.9. The 'IIIegitimate' Expressions 38

List of Figures XXI

6. I I/I Diagram traced by the Ambrym natives to explain their

LIST OF FIGURESkinship system 91

I 6.11/2 Lévi-Strauss's diagram to explain Aranda kinship inThe Savage Mind 91

frontis. Asmat shield XXIV 6. U/3 Malakulan sand-drawings 923.8.1/1 The index as the pivot ofthe art nexus 38 6.u/4 Nahal: 'the Path' 933.10/1 The self-made index 42 6.12/1 Performance art complementing graphic art: the Malakulan3.10/2 Tbe black disc loou as if it 'wanted' to retum to the centre dancer's path 94

of the square 43 6.12/2 'The Hawk flying from place to place' 943.10/3 Causality made visible: detail from P/uto and Persephone by

7.3/1 Volt sorcery and the paradoxical agency of representarion 103Bernini 44

••..2/1 Tbe prototype as agent: Samuel.1ohnson by Reynolds 537.5/1 The circularity of productive exchanges between index

(mauri) and prototype (hau) l<lq4.3/1 The multiple levels of agency within the index 54 7.6/1 Ti 'i in its benign form as a volcanic growth stone uo4.3/2 The hierarchical embeddedness of agent-patient relations 55 7.6/2 To'o with Oro the god hidden from view and wrapped4.4/1 Nail ferish figures from the Congo region of Africa 60 in bindings 112

4.4/2 The ferish as an index of cumulative agency and as the visible7.6/3 Sorcer)' ti'i with hollow back for feathers/exuviae U2

Imot tying together an invislble skein of spatio-temporalrelarions 61 7.7/1 Santoshi Ma II9

4·4/3 Mrs Pankhurst by Mary Richardson: the Rokeby Venus by 7.7/2 Ocular exchange as the medium for intersubjectivities 120

Velázquez slashed by Mary Richardson, 1914 63 7.U/I The fractal god: A'a from Rurutu 138

4·4/4 The shared biographical spaces of persons and images 65 7. I 1/2 Genealogical personhood objecrified: a staff god from the

5.~/i TwbrianJ l:anue pw\\-buanl 7° Cook Islands q8

6.2/1 latrnullime containers 757,U/3 'Vierge ouvrante' 142

6.3/1 The hierarchical strueture of the decora tive index 777.u/4 An Indian parallel to 'vierge ouveante' 143

6.4/1 The basic morions of pattem tormation 78 7.12/1 'World-seed': saiagrama 146

6.4/2 The linear reperirion of a single motif: simple pattem 78 7.12/2 Kumari with painted third eye 152

6.5/1 The multidimensional orientarion of a single motif: 8.7/1 Etua 171

complex pattem 80 8.7/2 Plane rotation in the formation of Steinen's two classes of etua 172

6·5# The pattem as mind trap: glide refleetion and 8·7/3 Refleetion and rotarion in the formation of etua L, M, and N6gure-ground reversal 80 in Fig. 8.7/1 173

6.Rjl Cdtic knotwurk: apotropaic pattem 84 8·7/ ••. Geometric orientation affccting the change of motifs:

6.9/1 Koiam threshold design: the pattern as topological snare 85 from etua to the female tattooing tortoise morif kea 174

6.9/2 The constiturive element of the ko/am 86 8.8/1 Etua O in Fig. 8.7/1 sliding by degrees to become the

6.10/1 Comparison of threshold, tattoo, and coin designs from hope vehine class of motifs 174

Layard 1937 87 8.8/2 From sitring etua to hope vehine 175

6. 10/2 The serial connection of points in the formation of 8.8/3 From hope vehine to vai o Kena 176the Cretan labyrinth design S9

I 8.9/1 Coordinate transformarion of a hope vehine motif resulting6.10/3 The Roman maze at Pula 89 in kake (type I) 176

XXIl List o[ Figures List o[ Figures XX1I1

8.9/2 Kake from hope vehine: duplication, rotation, and reflection 177 8.16/6 The rules ofbody proportions in Marquesan art 201

8·9/3 Coordinate transformation with rotation: from sitting etua 8.17/1 Marquesan artistic proliferation: fan handles divided intoto kake (type lI) 177 multiple segments in different orientations 203

8·9/4 The chiefess's girdle with type III kake 178 8.17/2 Double-anchor fan handles with morphologically adapted head 2048·9/5 Kea (tortoise) female tattoo motifs recorded by Steinen 179 8.17/3 Fly-whisk handle from the Austral'Islands with similar proboscis8.9/6 A tortoise of tortoises: kea motifs on tortoise-shaped terrine 179 face 204

8·9/7 'Realistic' kta tortoise tattoo recorded by Langsdorff 179 8.18/1 A simple form of stilt-step with rudimentary face at base 2058.10/1 Sitting elua hypostasized to mala Iroala 181 8. 18/2 From rudimentary face to complete second figure: stilt-step 206

8.10/2 Mala /roala in male tattoos recorded by Steinen 181 8.18/3 Transformation ofnose/mouth to proboscis 206

8.10/3 The mala Iroala tattoo indicated tends evidently towards 8.18/4 Stilt-step with multiple figures and 'mooning' buttocks 207the elua motif 181 8.18/5 Stilt-step with rotated duplicate figure 207

8.IO!4o From mata hoata to the Vau rock motifby the addition of spirals t~h 8.18/6 Maximally proliferated stilt-step with eight figures 20X8.11/1 From double mala "oala to ipu 182 8.19/1 U'u warrior's cluh with protective face proliferation 2098.11/2 Tapa mask with ipu 'eyes' 182 8.19/2 The figurative model for U'u displaying subsidiary figure8.11/3 'Hand face' design I 183 on the back 211

8.11/4 'Hand face' design 2 183 8.19/3 U'u exemplifying Steinen's medel with subsidiary head8.u/S 'Coiled shellfish' pcriri as a face motif 184 protecting the 'back of the back' 01'the club 2Ií

8.11/6 Langsdorff's 1813 engraving ofyoung warrior with eye and 8.20/1 Narrative art: 'Pahuatiti's daughters on their swing' depicted

face designs 185 on ear-plugs 212

8.12/1 Rosette motifs 186 8.20/2 Ear-plug with swing narrative moving towards the

8.1212 Hierarchization: niho 'teeth' motifs 187elua/mata heata motif 213

8.12/3 Anialiu band containing alternating reversed etua motif t 8.20/3 Ear-plugs depicting the obstetric myth displaying a fusion

in Fig. 8.7/1 187ofbodies 214

8.13/1 Figure-ground reversal and hierarchization: tortoiseshell 9.1/1 Malangan carving 224

incised disc 188 9·4/1 The artist's leuvre as a distributed object 235

8.13/2 From eyes to lizard: Steinen's evolutionary derivation from 9.4/2 A version of Husserl's diagram of time-consciousnessend-plate discs 189 from Gell, Tire Antlrropology o[ Time 239

8.14/1 From three dimensions to two in Western art: the back view 9.5/1 Tire Large Glass or Tire Bride slripped bare by lrer Baclrelorsof the image as the back view of the protolype 191 even by MareeI Duchamp 245

8.15/1 From two dimensions to three in Marquesan art: providing 9.5/2 Tire NetJlJork o[ Stoppages by Marcel Ducharnp 246

the imagc ,,;in a baek 192 9·5/3 The Young Matl and Girl i" Spring by Mareei Duchamp (1911) 2.•.88.15/2 From Lévi-Strauss 1963: split representation and the 9.6/1 The Maori meeting house as a collective index of

Janus-faced image in north-west American tribal art 194 communal power 252

8.16/1 Stone terraee of eeremonial platform 197 9.6/2 Kowhaiwhai patterns 254

8.16/2 Makii-taua-pepe: female divinity giving birth 197 9.6/3 The Mao'; meeting house as an object distributed in space

8.16/3 Wooden posts carved io two-dimensional form 198and time 255

8.16/4 The double tiki form in bone hair ornaments with

Ihakenkreuz motif on back, 2nd row 200 TABLE8.16/5 The Irakenkreuz motif: a variant of etua via hope vehine 200 I. The Art Nexus 29

7.1. Mimesis and Sorcery

I now turn to the anthropological theory of representational art, that is to say,the production, circulation, and practical use of indexes that have relevant pro-totypes, something that geometric patterns lacL Most Western art is of thistype, and most of art-theory is about representation, in one way or another. Ishall try to avoid, as far as possible, discussing Western art, tempting thoughit is to do so. Instead, this part of the argument pivots around the classically'anthropological' themes of sorcery and image··worship.

The basic thesis of this work, to recapitulale, is that works of art, images,icons, and the like have to be treated, in the context of an anthropologicaltheory, as person-like; that is, sources of, and targets for, social agency. In thiscontext, image-worship has a central place, since nowhere are images moreobviously treated as human persons than in the context of worship and cere-monies. In this chapter, I present a general thc:ory of idolatry, a prac!ice I byno means regard as more misguided than any other religious observance; andI am, I hope, capable of sustaining a sympathcric attitllde towards rcligion ingeneral, despite not being a religious person rnyself. 'Idolatry' has had a badpress since the rise to world domination of Christianity and Islam, whichhave both inherited the anti-imagistic strain {lI' biblical Judaism. Christianity,encllmbered with its Graeco-Roman inheritance, has had to struggle moreactively with recrudescences of de facto 'pagan' idolatry, and has experiencedcataclysmic episodes of iconoc1asm; Islam has more consistently resisted thelure of image-directed forms of worship. Protestantism, the most dynamicbranch of Christianity in the past few centuries, has hardly been less puritan-ical about idolatry than Islam, and the consciences of the Catholics have beenoften pricked on this account as well, so that lhe net result is today that 'ido 1-atry' is a pejorative word, which anthropologists, especially ones who c1aimto empathize with the religious sentiments of others, are not supposed to use.But rather than resort to some vague or misleading circumlocution I prefer tocall the practice of worshipping images by its true name, deeming it better toexylainidolatry, rather than rechristen it-by showing that it emanates, notfrom stupidity or superstition, but from the same fund of sympathy whichallows us to understand the human, non-artefactual, 'other' as a copresentbeing, endowed with awarencss, intcntions, aml passions akin to OUI" own.

Granted that idolatry may be so called, I still have to explain why a bookostensibly about 'art' has to devote so many pages to a topic which appears tobelong to the study of religion rather than aesthetics. Those who object to myanti-aesthctics stance will regard as irrelev:mt the 'religious function' of imagesthey prefer to sec in a totally non-religious frame of reference, from the stand-point of aesthctic contemplation. An image viewed as a source of religiouspower, salvation, exaltarion, is not appreciated for its 'beauty', but for quite dit:'ferent reasons, the proponents of 'aesthetic experiences' would argue. I regardthis as fallacio\!s on two grounds. First of ali, I cannot tell between religious -and aesthetic exaltation; art-Iovers, it seems to me, actually do worship images '> _~

in most of the relevant senses, and explain away their de facto idolatry by ration-alizing it as aesthetic awe. Thus, to write about art at ali is, in fact, to writeabout either religion, or the substitute for religion which those who have aban-doned the outward forms of received religions content themselves with. TheProtestant-Puritan heritage combined with a special form of art-theoreticalcasuistry have cstablished a special form of bad faith about the 'power ofimages' in the contemporary Western world, as Freedberg (1989) has pursuas-ively argued (scc below). We have neutralized our idols by rec1assifying themas art; but we perform obeisances before them every bit as deep as those of themost committed idolater before his wooden god-I specifically inc1ude myselfin this description. Secondly, from the anthropological point of view, we have .to recognize that die 'aesthetic attitude' is a specific historical product ofthe religious crisis of the Enlightenment and the rise of Western science, andthat it has no applicability to civilizations which have not internalized theEnlightenment as we have. In India (which figures largely in the ensuing dis-cussion) idolatry flourishes as a form of religiosity, and nobody in their rightminds would try to drive a wedge between the beautiful form and religiousfunction of venerated idols. In India aesthetics, as in the ancient world, is sub-sumed under the philosophy ofreligion, that is, moral philosophy, as a matterof course. Conscquently, it is only from a very parochial (blinkered) Westernpost-Enlightenment point of view that the separation between the beautifuland the holy, between religious experience and aesthetic experience, arises.Since this is so, the anthropologist writing about art inevitably contributesto the anthropology of religion, because the religious is-in some contexts,though not all----prior to the artistic.

Idols come in many varieties, but it is conventional to distinguish two polartypes; (i) purely 'aniconic' idols, such as the baitulia (black meteoritic stones)warshipped in ancient Greece, versus (ii) 'iconic' idols, that is, indexes physic-ally resembling a prototype, usually a human being, according to the formula[[Prototype-A ~ [Artist-A]] ~ Index-P, where the prototype is the god,whose 'likeness' is mediated by the artist. However, before starting, it may be aswel1to c1arifyour reasons for not paying very much attention to this distinction.Ali idols, I think, are 'iconic'--induding the so-called aniconic oncs--whether

or not they look like some familiar object, such as a human body. An aniconicidol is ~ 'realistic' representation of a god who either has no form (anywhere),or has an 'arbitrary' form, in the particular 'body' he inhabits for the purposesof being worshipped by his mortal devotees, here below. A meteoritic stoneis not a very, very, conventionalized or distorted 'portrait' of a god, who, elsc-where, looks Iike a human being. Dne need not imagine that worshippers ofstones would 'prefer' to worship more realisric porrrayals of their gods, buthave to make do with unshaped sranes for lack of any local stone-carvers of thenecessary ability. For such worshippers, the meteoritic stone is an index of thegod's spatio-temporal presence (and origins, in that the stone fell to earth fromheaven). The stone is a 'representative' of the god, Iike an ambassador, ratherthan a visual icon. Just as an ambassador can rt:present his country in Moscowone year, and the following year in Washington, the stone represents the godat whatever spatial coordinates it happens to bej and it is movable. Dne such'black stone' for instance, worshipped in Arabia as Cybele, was conveyed withgreat ceremony to Rome, where it was installed into a cavity within the idolof Magna Mater Idea, the goddess of members of the Imperial élite who fav-o'ured an 'internationalist' religious outlook (Dumézil 1980; on the significanceof such transactions, see below). The ideas of 'representing' (Iike a picture)and 'representing' (Iike an ambassador) are distinct, but none the less linked.An ambassador is a spatio-temporally detached fragment of his nation, whotravels abroad and with whom foreigners can speak, 'as ir' they were speakingto his national government. Although ambassadors are real persons, rhey area1so'fictions', like pictures, and their embassies are fictional mini-stares withinthe state; just as pictures show us landscapes and personages who are 'not reallythere'. AIthough the Chinese ambassador in London does not look like China,or the Chinese governrnent or people, he does have to be visible, and he doesvisib/y represent China on official occasions. He does not look like China, burin London, China looks like him.

Dne could not contrast a 'realistic' iconic ambassador to an 'unrealistic', ani-conic ambassador, and no more, perhaps, should one contrast 'realistic' idols,

, to unrealistic, aniconic idols. Whatever rhe idol looks like, that, in context, is: what the god looks like, so ali idols are equally realistic, because the idol-forrn

is the visual form of the god made present in the ido!. The contrast is notbetween idols which resemble human beings to a greater of lesser degree, butbetween gods who in idol-form (visually) resemble human beings to a greateror lesser degree. Idols, in other words, are nor depictions, not portraits, but(àrtefactual) bodies. The formula cited above thcn [[Prototype-A ~ [Artist-A] ]~ Index-P does not necessarily imply 'realistic' (i.e. anthropomorphic)depiction of the prototype by the artist in the form of the image. When sucha formula is applied to portraiture, this implication exists, because there isa Iiving model for the portrait, and fidelity to this living model is the socialobjective of portraiture. But in the context of idolatry, the agency involved is

religious. Where an idol is an artefact (rather than a natural obj.ect, such as ameteoritic stone) the nature of agency exerted by the prototype ISto cause theartist to produc~'a rdigious/y stipulattd image aceording to the conventi~ns ~orsuêh images, which may be iconic/anthropomorphic or abstract and anrconlc.ln either case, the artist has to produce a 'faithfuI' rendition of the features ofthe accepted image of the body of the god, triggering 'recognition' of the godamong his worshippers.

To assert that in the context of idolatry, the idol is not a 'depiction' of thegod, but the body of the god in artefact-form is ali very well, but I ac~ept thatáilYsuch assertion constitutes a paradox, and 1must labour hard to dlspel thepuzzlement it must inevitably produce in the mind of any reasonable person.The mystery 01' the animation of idols, their genuine, if peculiar, personhood,has to be appro:lched, Iike any difficult problem, via a series of incrementalsteps, not to mention detours through some unfamili~r territory. Rath~r thanõea! with idolatry in its most elaborated forrn, 1shallllltroduce the subJect bypresenting an analysis of yolt sorcery {énvoutement)-the practice of inflictingrÚlrmon others via theirimages> This is a particularly pertinent case of agencymediated via representational indexes, and it has the advantage of removing thediscussion, temporarily, from the sphere of 'religion proper', which involvesthe always problematic question of other-worldly entities such as gods. Voltsorcery takes place in this world and without the neccssity for ot.her-worldly<Iivineor diabolic intervention, though it may be sought. As we wIIIbe able toobserve there is actually a smooth transition between this-worldly volt sorceryand religious image-worship orientated towards other-worldly beings. Thiswill be explaincd in due course.

7.2. The Mitnetic FacultyFrazer rather than Tylor is the ancestral figure who presides over this dis-cussio~. Frazer, it will be ~ecalled, distinguished two basic modes of 'magical'action; (i) contagious magic, the magic of c?fltact~ in which influence passesfrom one object to another, and (ii) sympathetic magic, which depends on~~_properties, that is, if object A shar~s p~opertie~.w,ith,obJect B, A h~smfluence over B 01' vice versa. The Frazerlan Idea of Imltatlve sympathetlcmagic has had an enormous influence on aesthetics and the philosophy of artin the course of this century. Via Benjamin and Adorno, this influence seemsset to continue; Taussig (1993) in a most exciting recent book, whose primaryinspiration comes from Benjamin, has this to say about Frazer:

I am particularly taken by [Frazer's] proposition that the principie underlying the imi-tative component of sympathetic magic is that 'the magician infers that he can produceany efTect he desires merely by imitating it', (52) Leaving asi~e for ~he moment thethorny issue of how and with what success Frazer could put hlmself mto the head ~fone of these magicians, and to what degree either the accuracy or usefulness of hls

propositiondepends on such ,(move,Iwant to dwellon this notion ofthe eopy, in mag-ieal praetice, affieting the original to sueh a degree that the represt?ntation shares in oraequires the properties ofthe represenred. To me this is a disturbing notion, foreignandfascinatingnot becauseit so flagrantlycontradicts the world about me but rather, thatonce posited, I suspect if not its presence, then intimations thereof, in the strangelyfamiliarcommonplaceand unconscioushabits of representationin the worldabout me.(Taussig 1993: 47-8)

Following Benjamin, Taussig bases his analysis of mimesis in the colonialmilieu (ofthe CUna Indians) Ona supposed 'mimetic faculty'. Benjamin thoughtthat rhe mimetic faculty, which in modem times has resulted in a world filled tooverflowing with images and simulacra, so that nothing seems real any more, hadits origin in a primitive compulsion to imitate, and thus gain access, to the world:

[Man's] gift of seeing resemblancesis nothing other than a rudimenr af the powerfulçompulsionin former times to becomeand behave Iikesomethingelse. Perhaps thereis none of his higher functions in which his mimetic faculty does not play a deeisiverole. (1933; cited in Taussig 1993: 19)

Taussig's book bears witness to the productivity of Frazer's idea mediatedthrough Benjamin's surrealist imagination, but it is fair to say that the 'mimetkfaculty' is only rather vaguely delineated. The facl that 50 much human behaviouris imitative does not necessarily imply the existence of a 'faculty' inherited fromthe distant pastoAímost allleamed behaviour could be described as imitat.ive in

, _'- __ .'.'-" _"0_- .' .' ,

é· that it is based on the imitation of an intemalized model. Mimesis, narrowlydefined, involves the actual production of images (indexes) whose salient prop-erty is prototype, via resemblance to the original, and within this category arte-facts, having visual resemblance to the originaIs, can be acrorded a separate status.

What Frazer never explained is why the mutual resemblance of the imageand the original should be a conduit for mutual influence or agency. He attrib-uted it to a mistaken hypothesis, akin to a scientific theory, but grounded inerror. The trouble is that if the practitioners of sympathetic magic could haveseen their practices as Frazer saw them, they would never have engaged in themin the first place. By abstracting a generalizable 'principIe' from the inchoateworld of practice, Frazer guaranteed his eventual misunderstanding of the datahe had at his disposa!. Taussig, in his Benjaminesque reanalysis, argues thatthe basis of sympathetic magic is not a tragic misunderstanding of the natureaf physical causality, but a consequence of epistemic awareness itself. To see(or to know) is to be sensuously filled with that which is perceived, yieldingto it, mirroring it-and hence imitating it bodily (ibid. 45) (see above in'Captivation', Sect. 5.2). But for the moment I will approach the sympatheticmagic problem from a different direction.

Frazer's intellectualism treated magic as a I()rm of mistaken causal think-ing. Anti-Frazerians, ever since, have criticized Frazer for attributing 'causal'

intentions to behaviours which were symbolic or expressive (Beattie 1(}66).There is another approach, though, which can be adopted, which is not to ron-demn Frazer for having invoked causality at ali (because magic is, after ali,intended to cause things to happen) but to rethink the idea of'cause'. Frazer'smistake was to impose a pseudo-scientific notion of physical cause and effect(encompassing the entire universe) on practices which depend on intentional-ity and purpose, which is precisely what is missing from scientific determin-ism. Magic is possible because intentions cause events to happen in the vicinity ofagents, but this is a different species of causation from the kind of causationinvolved in the rising and setting of the sun, or the falling of Newton's appleetc. For instance: here before me is thisboiled egg.What has caused this eggto be boiled? Clearly, there are two quitedifferent answers to this-(i) becauseit was heated in a saucepan of water over a gas-flame, or (ii) because I, off myown bat, chose to bestir myself, take the egg from its box, fill the saucepan,light the gas, and boi! the egg, because I wanted breakfast. From any practicalpoint of view, type-(ii) 'causes' of eggs being boiled are infinitely more sali-ent than type-(i) causes. If there were no breakfast-desiring agents Iike meabout, there would be no hens' eggs (except in the South-East Asian jungle), nosaucepans, no gas appliances, and the whole egg-boiling phenomenon wouldnever transpire and never need to be physically explained.~o, whatever theverdict of physics, the real causal explanation for why there are any boiled eggsis that I, and other breakfasters, intend that boiled eggs should existo. There is nothing mystical involved in tracing the causation of events in one'svicfnity to intentions or acts of willing or wishing performed by oneself or otheragents in one's neighbourhood. That is how perfectly ordinary events do ordin-arily happen-barring 'accidents', and who is ever to know that an accident hasnot been willed by somebody? Frazer's mistake was, so to speak, to imaginethat magicians had some non-standard physical theory, whereas the truth isthat 'magic' is what you have when you do wíthout a physical theory On thegrounds of its redundancy, relying on the idea, which is perfectly practicable,that the explanation of any given event (especially if socially salient) is that itis caused intentionally.

The causal arrow between desire and accomplishment reflects the practicalfact that the more one desires something to happen, the more Iikely it is tohappen (though it still may not). Magic registers and publicizes the strengthofdesire, increasing the (inductively supported) Iikelihood that the much-desired, emphatically expressed, outcome will transpire, as frequently happenswith respect to those outeomes we loudly c1amour for. 'Ali events happenbecause they are intended'-'I emphatically intend that X shall happen' trgo'X shall come to pass'. This is not 'confused' physics, nor is it devoid of a basisin social experience, as Malinowski (1935) understood more clearly than any ofhis successors, with the possible exception ofTambiah (1985).

7.3. Volt Sorcerj!The symbolic language developed in earlier scctions of this essay c-anbe ap-plied to the kind of causality which is involved in Frazer's prime exampie ofsympathetic (imitative) magic, which is also extensively discussed by Taussig,riamely, that form of sorcery in which an imagc of the victim is made \often (lfwax or some vulnerable material), subjected to injury or destruction, with theresu!t that the victim of the sorcery suffers the same injuries or is done awaywith entirely. This kind of sorcery is practised in innumerable forms, ali overthe world, as anyone who opens Frazer's book is able to discover.

'[his kind of sorcery can be practised in aniconic forms; the ancient Greeks,for instance, used to sorcerize one another by taking small pieces of lead, onwhich they would scratch the name ofthe victim, and the words 'I bind, I bind'before burying them in the ground. This would cause the victim to sicken anddie. The comparison with volt sorcery, which uses the image of the victim inplace of the inscribed lead is interesting, in that it suggests that 'visual repres-e!1tation'and 'binding' (and naming) are essentially akin to one another. 'Binding'is, indeed, a fundamental metaphor of magico-religious control which we willencounter below. The words 'I bind, I bind' provide a bridge between the lin-guistic side ofmagic (they are words) and the physical side ofmagic (they referto a physical operation, carried out on the body of the victim). The action oftp.akinga representational image, of any kind, involves a kind ofbinding, in that.!he image ofthe prototype is bound to, or fixed and imprisoned within, the index.

N'e!vousness about being represented in an index (a photograph or portrait)is often discussed as if it were only a foible of innocent tribesmen, who believethat their souls are in danger of being stolen away therein. In fact, almosteveryone has reasons for wishing to keep some degree of control over repres-entarions of themselves, rather than have them circulate freely. I might beresigned to having my face photographed and circulated, but I do not feel thesam~ about my naked behind. There is no reason to invoke magical or animisticbeliefs in order tosubstantiate the idea that persons are very vulnerable indeedto hostile representation via images, not just to cruel caricatures, but even viaperfectly neutral portrayals, if these are treated with contumely or ridicule. Itis not just that the person represented in an image is 'identified' with thatimage via a purely symbolic or conventionallinkage; rather, it is because the~gency ofthe person represented is actually impressed on the representation. Iam the cause of the form that my representation takes, I am responsible for it.I cannot disown a photograph ofmy inelegant posterior, on the grounds that Idid not press the shutter and cause this damaging image to come into existence.I can blame the photographer for taking the picture, but I cannot blame himfor the way the picture carne out.

The 'magical' aspect of volt sorcery is only an epiphenomenon of our failureto identify sufficiently with sorcerers and their victims, our estrangement from

~Index·A(volt) h •. Prototype-P

~u~s arm to (vlctlm)

Sorcerer-A Index·P(volt)~injureS ••

$orcerer"A(artlst) Index-P(volt) Prototype- Reclplent~akes"

Prototype-A(virtlm) Sorcerer-P(art/st)••FIG. j.3/I. Volt sorccry and the parodoxical agency of representation: the victim is both agentand patient

them, not the result of their enslavement by superstitious beliefs entirely dif-ferent from our own. We suffer, as patients, from forms of agency mediated viaimages of ourselves, because, as social persons, we are present, not just in oursingular bodies, but in everything in our surroundings which bears witnessiô-õue existence, our attributes, and our agency. Volt sorcery is not a moremagical but just a more literal-minded exploitatiori: of the predicament ofrepresentability iniÍnage formo It does not take leave of the everyday world,lnappealing to some occult force, some magical principie of causation; on thecontrary, it unites cause and effect ali too c1osely, so that the causal nexuslinkillg the image to the person represented is made reversible-the image canexercise a causal effect, in the opposite direction, over the person. The modusoperandi of volt sorcery can readily be expressed using the diagrammatic con-ventions introduced earlier, as in Fig 7.3/1, or:

[[ [Prototypc-A] ~ Artist-A] ~ Index-A] ---? Prototype/Recipient-P.The verisimilitude, so to speak, of volt sorcery resides in the fact that the

victim appears twice; once as the prototype who causes the index to assume itsparticular form, and once as the recipient, whose injuries stem from the injuriesthat the index has received. The victim is ultimately the victim of his ownagency, by a circuitous causal pathway. Vulnerability stems from the bare pos-sibility of representation, which cannot be avoided. Sorcery beliefs endure, andare highly explanatory, because vulnerability to sorcery is the unintended con-sequence of the diffusion of the person into the milieu, via a thousand causalinfluences and pathways, not ali of which can be monitored and controlled.Prazer himself noted that image-based sorcery is c10selyallied to the other kindo(sorcery which depeÍlds on exuviae; hair, nail-clippings, food leftovers, ex-ereta, and the like. Often, the volt is rendered more effective by the incorpora-tion into it of the victim's exuviae, ~othat imitative magic, based on (visual)similarity, is allied with the other main kind of magic, based on contact. Onceagain, to describc this as 'magic' and to imagine that such sorcery is based on

occult principIes (or symbolic rituaIs) is quite misleading. There is nothingtranscendental about the kind of causality involved in exuviae sorcery, thoughoften sorcerers may enlist spiritual helpers in pursuing the objects of thórhatred. Exuviae sorcery works (or seems likcly to work) because of the intim-ate causal nexus between exuviae and the person responsible for them. Theseexuviae do not stand metonymically for the victim; they are physically detachcdfragments of the victim's 'distributed personhood'-that is, personhood dis-tributed in the milieu, beyond the body-boundary.

The interest of exuviae sorcery, from our point of view, is that it forges adirect link between the index as an image or the prototype, and the index as a(detached) part of the prototype. We are not accustomed to think of images(such as portraits, etc.) as parts of persons, limbs, as it were. In terms of thesemiotic theory of representation, nothing would be more erroncous than toimagine that the substance of a sign (the visible 01' audible sign 'dog') were partof any dog, 01' dogs in general. But with indexes it is not the same as withproper signs. Abduction from an index does characteristically involve positingá substantive part-whole (01' part-part) relation. Smoke is a kind of 'part' offire, for instance. A person's smile (the cheshire cat cxcepted) is a part of lhefr!endly person it betokens. From this point of view, it is not senseless tosuppose that Constable's picture of Salisbury cathedral is a part of Salisburycathedral. It is, what we would call, a 'spin-off' of Salisbury cathedral.

The convergence between images of things and parts of things can heapproached from a philosophical angle. Yrjô Hirn (1900, cited by Frazer 19Hoand discussed also by Taussig 1993: 51) made the suggestion that the magicof similarity and the magic of contact were really one and the same, because'primitive' people anticipated, in their confusion, the philosophical doctrine of'emanations'. Hirn writes:

For it is evident that a philosophieal doetrine, if it fits in with the facts of primitivesuperstition, may be explanatory of those vague and latent notions whieh, without logicaljustifieation 01' systematieal arrangement, lie in the mind of the magician and the idol-ateI'. Sueh a doetrine is presented to us in the familiar emanation-theories, aeeording towhieh every image of a thing eonstitutes a eoneTl~tepart of that thing itself. Aceoroingto the dear and systematie statement of this ooetrine given by the 010 Epieureanphilosophers [1wil/ provide the relevam quotationfrom Lucretius below] shadows, reflec-fions in a mirror, visions, and even mental representations of distant objeets, are ;tIleaused by thin membranes, which eontinually d<:taeh themselves from the surfaees ofali bodies and move onwards in ali direetions through spaee. lI' there are sueh things asneeessary miseonceptions, this is eertainly one. Sueh general faets of sensuous experi-enee as refleetion, shadow, and mirage will naturaIly appear as the resutt 01' a purely

material decortkation-as in a transfer picture. How near at hand this theory may lieeven to the modern mind appears I'tom the eurious fact that such a man as Balue relIbaek on it when attempting to explain the newly invented Daguerrotype, that most mar-veIlous ofimage-phenomena. (Hirn 1900: 293-4; cf. onphotographs as indexes orthereal presenee 01' persons, Barthes 1981 cited and discussed in Freedberg 1989: ch. 15)

The doctrine of emanations comes directly from Epicurus, but the mostfamou;;; and for my purposes the most interesting, statement of the doctrine isprovided by Lucretius, who writes:

I will attempt to lay before you a truth which most eoncerns ... the existence of mostthings we calI the idols [simulacra; in Greek, eidola] 01' things: these, like lilms peeledoff from the surface of things, fly to and fro through the air ... I say then that pieturcsaf ltlings aua ihin shapes are emitted from things off their surrace, to which an imageserves as a kind of film, 01' name it ir you like a rind, because such image bears anappearance and form like to the thing whatever it is from whose body it is shed andwanders forth.This you may learn however dulI of apprehension from what folIows.[Many visible objects], ... emit booies some in a state 01' loose diffusion, Iike smokewhich logs 01' oak, heat and fires emiti some of a doseI' and denser texture, Iike the gos-samer coats which at times eieades doff in summer, and the films which calves at theirbirth cast from the surfaee of their body, as welI as the vesture whieh the slippery ser-pent puts off among the thorns; for often we see the brambles enriehed with their flyingspoils: since these eases oeeur, a thin image likewise must be emitted from things offtheir surfaee. (De Rerum Natura 4: 26 ff. trans. Munro, pp. 44- 5)

Lucretius attributes the flying siIl1ulacra of things to a kind of internal ;ostl-ing within objects, which causes the minute bodies 'in the front rank' to bedischarged from the surface and to fly outwards. The simulacra are physicalthings though, and we see objects because simulacra enter our eyes and we canfee/ them 'since a particular figure felt by the hands in the dark is known to bethe same which is seen by the bright Iight of day, touch and sight must beexcited by quite similar causes'. Lucretius discusses a number of optical phe-nomena, notably images reflected by mirrors, but most interesting, perhaps, isthe way in which he consistently draws analogies between vision by means of'idols' and other physical forms of diffusion into the ambience, particularlysmell and smoke, as well as the shedding of skins, rinds, and films from thesurfaces of things. I shall have occasion, later, to return to the conceptualIinkage between smoke, smell, skins, and visual appearances, which ~üchlerhas identified as key elements in the ideology associated with a particularlywell-known Melanesian art form, the Malangan carvings from northern NewBritain (see below, Sect. 9.2). For the moment, though, I am interested in Him'spoint that if 'appearances' of things are material parts of things, then the kindaf leverage which one obtains over a person 01' thing by having access to theirimage is comparable, 01' really identical, to the leverage which can be obtainedby having access to some physical part of them; especiaIly if we introduce the

notion that persons may be 'distributed', Le. ali their 'parts' are not physicallyattached, but are distributed around the ambience, like the discarded 'gossamercoats of cicadas' in Lucretius' memorable instance, which are both images andparts of the living creature.

Babadzan explains that the exchange process dcscribed in Ranapiri's textinvoh'es three participants, the priests (lohunga), the 'hau ofthe forest', and thehunters. The 'hau of the forest' can be glossed as the 'principIe of increase' infile forest; its fertility in other words. The priests make offerings of mauri tothe hau of the forest. The hau of the forest responds by providing the hunterswith oirds to capture. A portion of these birds must be returned to the priests.The ofTering made by the priests and referred to as mauri are 'fertility stones'.The mauri forge the link between Ranapiri's famous text and our theme in thischapter, for maur; are aniconic idols. They are indexes, in other words: they areobjectified repõsitorIes ofihe spirit (ofincrease) ofthe forest. They might justoeS"pecialst6ries,butBest also tells us that sometimes they took the forro of'a hollow stone, in which holIow would be placed a lock of hair or some otheritem. These articles would be deposited at the base of a tree, or hidden in ahole, or by the side of a tree' (ibid. 438). Another form of mauri was createdby immuring a (Iiving) lizard within a holIow tree beside a bird ...snaring site(ibid. 437). The significance of the hollowness of mauri idols will emerge in duecourse (below, Sect. 7.II), '

Let l1Sreturn to Babadzan's exposition of the ritual sequence, in which, inexchange for the birds they captured in the forest, Maori hunters were obligedto recompense the priests (tohunga) with a portion of the game they secured.

I turn now to an ethnographic example which provides a bridge between voltsorcery, surely a discreditable practice, and the worship ofimages in the religiouscontextoVolt sorcery provides a model for undcrstanding the worship of imagesin generàl~indeed for 'objectification' inreligious contextsgenerally. The

màterial-and the idea-comes from Alain Babadzan's remarkable reanalysisof the religious practices of ancient Tahiti, and'thedeployment of images therein.-, -It may have been noticed that the victim of volt sorcery is involved in aninvoluntary process of exchange. Thisârises naturally from the fact that heappears twice in the formula, once as the contributor of something (his appear-âÍlce) and once as the recipient of something (injuries matching those sufferedby the volt, the index). He is an 'involuntary' agent; voluntary agency lies withthe sorcerer-who may, of course, have been jl1stifiablyprovoked.

The originality of Babadzan's account of Polynesian sorcery and idolatryarises from the subtle way in which he has observed both of these were varia-tions on the well-known, but often tantalizing, explanation of the process ofexchange provided by Ranapiri, the Maori intellectual, to Elsdon Best (cf.Mauss 1954; Sahlins 1974)·

I will now speakof the hau, and the ceremonyof whanga; hau. That hau is not the hau(wind) that blows-not at ali. I will carefullyexplainto you. Suppose that you possessa certain article, and that you give that article to me, without price. We make no bar-gainover it. Now, I givethat articleto a third person, who,after sometime has elapsed,decidesto makesomereturn for it, and so he makesme a present of somearticle. Now,that article he givesme is the hau of the article I first receivedfrom you and then gaveto him. The goodsthat I receivedfor that item I must hand over to you. It would notbe right for me to keep such goods for myself, whether they be desirable items orotherwise.I must hand them over to you, becausethcy are a hau ofthe article you gaveme ...

I will explain something to you about the forest hau. The mauri was placed orimplanted in the forest by the tohunga. It is the maut; that causesbirds to be abundantin the forest, that they be slain and taken by manoThese birds are the property of,belongto, the maur; ... Hence it is said that offerinl~sshouldbe made to the hau of theforest. The tohunga (priests, Rdepts)eat the offeringbecausethe maur; is theirs: it wasthev who located it in the forest, who caused it to be. That is why some of the birdscooied at the sacred fire are set apart to be eaten by the priests only, in order thatthe huu of the forest-product:"and the mauri, may return again to the forest-that is,to the maur;. (Best 1909: 439)

The magicstonesand the birds the hunters capture are thus, ritually,oneand the samething, one and the same toanga (gift) that the priests 'give' to the hunters, via the inter-mediaryagencyof lhe (orest. (Babadzan1993: 64, my trans.)

In other words, lhe forest is obliged to give birds to the hunters, because its life,its capacity to manifest productivity and fertility, is not its own; it has beenplaced therc by the priests. 'The hau of the forest, which meanwhile is, inMaori theory, the very principIe of the productivity of the forest, is thus con-sidered as the passive agent in a transmission of which the priesls are lhe primemovers' (ibid., italics in original, my translation). But there is a paradox here.How can a forest, or anything else for that matter, be a 'passive agent'-in asense, this is a contradiction in terms, as passivity is defined as the absence ofagency, and vice "ersa. What is meant is that the forest is passive in relation tothe priests, the prime movers, but not that it has no intrinsic agency at alI.There can only be a mauri of the forest, a physical objectification of the pro-ductivity of the j()rest, because the forest is (potentially) productive in itself; itis this 'agency' which has been co-opted by the agency of the priests.

The fundamental similarity between the situation just described and voltsorcery will perhaps be beginning to show itself: the mauri (objectified fertil-ity) created by the priests and buried in the forest is both a representation ofthe productivity ofthe forest (ex ante) and a cause ofthe forest being fertile (expost). Because ex ante fertility can be represented-i.e. objectified in an index-it comes under the control of those who control the index, the priests. The

priests, in other words, make an index ofthe ptoductivity ofthe forest and thatmakes the forest productive. They are rewardcd by the hunters, who receivefrom the forest the fruit of its productivity, and who return it to the priest~.The magic, so to speak, just has to work, because the index of productivity iscaused by the productivity it causes; there is a perfect, but invisible, circularity.The objectification of the productivity of the forest differs from volt sorcery,of course, in one essential respect, namely, that the maur; represents the forestas prosperous and flourishing, whereas the volt represents the sorcery victimas injured, or even dead. But the mechanics are the same; the priests makethe forest flourish by representing it as flourishing; on the other hand, thevery same objectification could be used to injure and kill the forest, were itsubjected to abuse. In fact, if enemies managed to find the mau'; of the forest,they would destroy its effectiveness by reciting productivity-negating chantsover it, bringing death to the owners, for now these same mau'; would object-ify misery, and cause it. This is the equivalent of sticking pins into a volt toinjure the victim represented as injured.

What is most interesting of ali to note, though, in this connection (Babadzan1993: 61) is that hau (fertility-principle) objectified in the mauri, is also theword used to referto the exuviae used by a sorcerer to ensorcell his victim(citing Tregear 1891: 52). We now see why holIow mauri had 'Iocks of hair'placed inside them. Babadzan suggests a most satisfactory explanation of thesyncnymy between sorcery exuviae and the principIe of fertility: both involvegrowth. Exuviae are paits of the body which have grown and become separate;this particularly applies to hair and nail-c1ippings; indeed, where adults areconcerned, these are the most obvious manifestations of growth that the humanbody provides; and even if hair remains uncut, it still falls out and separatesitself. Exuviae represent 'growth' because they are, so to speak, continualIy'harvestedi from the living body. Just as we harvest our hau, whenever we haveallaircut, so when the hunters enter the forest to 'harvest' the birds there, theyare harvesting the exuviae (hau) of the forest.

Exuviae soreery is possible because of the faet that as the body grows, itsheds its parts, and these beeome distributed around the ambienee. Here onemay reealI the remarkable resonance between Lucretius' instaneing the skinsshed by snakes and eicadas as prototypical sinlulacra or 'idols'; for these are,preCisely, exuviae produced by growth and distributed around. In faet, theEpicurean theory saw the generation of simulacra as a 'growth' process-the"shedding' of ephemeral skins from ali things indueed by a kind of 'pushing'from within.

The mauri, in its guise as a fertility stone rather than as an item of humanexuviae, objectifies the growth of the forest because it is, eonceptualIy, some-thing produced by that growth; it is the exuviae of the forest, which fullsinto the hands of the priests, who use it in lhe 'white' magie of prosperity-induetion rather than the black magie of sorcery and dearth-induetion.

Prlests mllu,; •..._T_hé_hII_U_--..J ~ EJunters(create mau';) -+ (productivlty) _of the Forest

'-- --..J

t hau of maur;

FIG.7.S/1. Thecircularity ofproductivc cxchangesbetween index(mauri) andprototype (hau) afterBahadzan 1993

t hau of birds

Physical1ysp~aking, .mau'; fertility stones could take a variety of forms, as wehave seen. Fórest maur; were anieonic, but those used to ptomote the growthof swe~t pot~ioes were earved às anthropormorphic images with open mouths,sometinies with their hands in a põsition suggesting the plaeing of food in themouth. Ensuring that the poeple ate, they were shown eating-and they wereof course fed themselves, with offerings offirst-fruits and so on. Fig. 7.5/1 is aversion of the figure Babadzan provides to summarize his general argumento

7.6. Decorticatz'onand the Exchange of Indexes: Tahitian To'oThe Tahitian equivalent of the Maori mauri stone took a variety of forms also.The produetivity offisheries was eontrolIed by a type of ti'i (tiki = 'image')calIed púna. These 'fish' stones were elongated, Iike fish, and eertain exampleshave fish-like gill slitson them. There are descriptions of the rituais associatedwith puna which make it clear that their use was entirely analogous to mau';.For agricultural purposes, Babadzan convincingly argues, a different type ofstonewas used, carved with anthropomorphie features, like the anthropomor-phic mauri buried in sweet-potato fields in New Zealand (ibid. 75-82). Thesegrowth stones wcre actually believed to grow, though only very slowly. Notmuêh documentation that is relevant to agricultural rites in Tahiti has unfor-tunately survived, but it is safe to say that increase-rituals comparable to themuch better docmnented Maori ones were praetised there (ethnohistorically,Maori civilization and language have antecedents in the Society Islands, ofwhich Tahiti is the most important). Perhaps the most important point to bearin mind is that the Tahitian word tupu, which basicaíly means 'growth' alsomeans, like hau, exuviae, usable for sorcery purposes. Tupu is the Tahitianequivalent of hau, in other words.

The interest of the Tahitian material discussed by Babadzan Iies in a differ-ent direetion than the mere replication of the Maori pattern of agricultural ritesand the eoncept~ of lIau, mauri, etc. Maori society was effectively decentral-ized throughout, but in ancient Tahiti there was a proto-state with a centrali:redcult, the cult ofOro.Consequently, in relation to the Tahitian material, we areable, as it were, to pass from volt sorcery (a private affair) to public worship ofthe most majestic and awe-inspiring variety, involving the state god, the 'godofpower', Oro.

FIG.7.6!1. Ti 'i in its benigll form <IS

a volcanic growth stone. Source: afterPierre volcanique 30 em., SocietyIslands. Musée de I'Homme, Paris

via representation, just like the forest hau in the previous example, or the pas-sive victim of volt sorcery.

Pursuing this theme, we are able to discover that the primary ritual pro-cedure in which the to 'o were involved, was a rituaI"of decortication, of thtcontrolIed production of divine exuviae. FollowiniBabadzan's analysis of theexisting sources, I will now describe this ritual sequence.

The primar)' to 'o were embodiments ofOro. The Tahitian political systemrevolved around asserting c~~.t~o!,o"er the marae (temples) of Oro; whoevercontrolIed the to 'o, controlIed the country, and rival chiefs fought bitterly overthem. The important to'o stood at the pinnacle, so to speak, of a polity ofimages, and social rank among human beings was strictIy and precisely corre-lated with the 'rank' of the to 'o (and lower down the scale, ti'i, 'sorcery idols')in their charge. S~l:ial rankandpolitical power were given regular and formalexpressio~llt aritual called pa 'iatua, which, translated, means the 'wrapping ofthe gods' (pa 'i = wrap up, tie, bundle). (On wrapping in general see Gell 1993.)-- Only during this ritual were to 'o revealed to view, and even then, only tothe most important and ritualIy protected priests and chiefs; the sight of ato 'ú would instantly result in the death of a lesser person, so filled were theywith mana (a quality called ra 'a in Tahitian). To protect them from view, to 'owere âtail other times tightly wrapped in bindings of sennit cordage and tapabark-c1oth, som(:times with roughly delineated facial features and limbs wovenõnto the outside (Fig. 7.6/2). The more important ones were kept in specialcontainers for storage and transporto The ritual of pa 'iatua consisted of theassembly of alI the images in a district at the marae, for the purpose of therenewal of their outer wrappings, especially the outer wrapping of the primarytó'o ofOro.

I need not describe the ceremony in any detai!. Essentially, it consisted of aprocession ofthc gods into the marae, led by the main god (completely hiddenin its ark), with lhe lesser ones following in order of precedence, followed (andprotected) by 'sorcerers' bearing not to 'o, but ti'i, sorcery images (Fig. 7.6/3).Lesser gods and ti'i 'sorcery' images, Ellis tells us (and there is a sole surviv-ing example in the British Museum to back this up) were not in the aniconic'pillar' style, but were hollow anthropomorphic images, with an internal cavityfor holding sacrcd feathers and/or sorcery exuviae (Ellis 1831: i. 339). Aliwould assemble in and around the marae, according to their station. In a separ-ate sanctuary, well away from public view, the to 'o would be divested of theirwrappings, oiled, and laid out in the sun. There then ensued an importamexchange sequence.

The most symbolicaIly significant 'currency' of political authority in Tahititook the form ofthe f~athers, the most important being red ones. Wearing redfeathers was a sumptuary privilege of the highest chiefs, the maro ura (redfeather girdle) was synonymous with occupancy of the paramount chiefship.

Unlike ti 'i images (see Fig. 7.6/1 )-such as fish stones and anthropomorphiccarvings, used in agriculture and sorcery-the objects at the centre of the Orocult, and the cults of lesser gods, ancestors of important kin-groups, were notanthropomorphic, though they were (in a sense) representational, and theywent by a different nameito 'o. The word to 'o means a 'prop' or staff 01' pillar.The to 'o represent, mythologically, the pillar placed bythe creator god to holdup the sky and preserve the ao (theworld of light and ofhuman life) from theericompassing powers of night, darkness, and divinity (the po). It is a mootpoint whether the to 'o deserve to be considered representational (iconic) ornon-representational (aniconic). As elongated bilIets of wood, slightly thicktrat one end than at the oiher, they might be regarded as 'realistic representa-tions' of pillars or props, which is what the)' are called. One could call them'iconic' indexes of pillam, in that they refer to the mythological pillars which!ío1d up the sky without actually being thest: pillars. On the other hand, theyrepresent, aniconically, gods who have anthropomorphic attributes. Oro wasnot a pillar, as such. Rather, as the pilIar which keeps heaven and earth apart--·and thus connects them--the to 'o is invested with the god's presence by virtueof contiguity rather than resemblance. Babadzan himself suggests that the to 'ois formless because the god presides over the origination of everything thatpossesses form; as the giver of forms to ali things, the god is himself form!ess.This deduction is well supported in the Tahitian creation chants (Henry 1928;cf. Gell 1993: 124 ff.). In this case, the to 'o is again an iconic imagc, a formles,image or formlessness, 01' ali things in statu nascendi. Here, it seems to me th:ltthe supposed contrast between iconic and aniconic representation comprehens-ively breaks down. The to 'o are wholly iconic and wholly aniconic at the sametime. What is quite inarguable though is that they have a prototypc, and thattheir prototype is the person of the god. Tht:y are also, like the humblest volt,open to manipulation by human beings. Through them, Oro himself, themightiest being in the Tahitian universe, could be rendered a 'passive agent'

FIG.7.6/z. To 'o with Orothe god hidden from viewand wrapped in bindings.Source: Museum ofArchaeology andAnthropology, Cambridge,Z 6067

When ali were sealed, the high priest opened the ark and took out the dreaded image,and as he uneovel'cd it upon the mat, the others ali uneovered theirs in unison ... Theminor gods then exposed, with their wrappers folded under them, remained in thehânrls of their owners, facing the 'ava 'a ready for presentation to the tutelar god [í.e.Oro] . . . When the image of the tutelar god was revealed from the profusion of redand yellow feathers lying upon its many coverings ... Then followed the presenrationof the minor gods by their owners in their proper turns, wíth offerings of new ura[red feather] amulets and loose feathers, which were given through the high priestto the tutelar god in exehange for some in his possession [i.e. whieh had iust beenunwrapped from the sennit wrapper of Oro]. This aet was called taritoara 'a-atua(the god's exehange) and was supposed to add new power from the greater god to thelesser ones. The tishermen's gods were presented last beeause they were from the sea.(Henry 1928: 166"7)

Babadzan provides the foIlowing schema of the (eather exchange: new, non-sarred feathers passed from the lesser priests to the priests of the primary 10 'o(the 'main' Oro) and the feathers which had previously been in contact withthe primary to 'o were passed back to the inferior 10 '0. In this way a portion ofthe sanctity of the primary 10'0 was distributed to the assembly of lesser 10'0 ineiênange for a tribute paid in the form of new feathers, a 'natural' product, asif\vere, potcntially embodying power and fertility, but not yet able to generateit. Only feathers which, as Babadzan notes, had becn intimately in eontaet withthe primary image of Oro for some lime, had this generative property (1993:116). The fcathcrs had to live and die with the god,quite literaIly, because itseems that to beeome divine exuviae, the primary god has to 'die' and scatterexuviae back into the world. This, Babadzan thinks, is the signifieanee of thesequence as a whole, that is to say the unwrapping of the gods foIlowed bywrapping them up again. Unwrapping brings the god into the world, wrappinghim up again scnds the gõd baek to the nether world, the po. But because thegod has died but left something behind (his feathers, his exuviae IUpu) thep()wêi- of the god is also left behind-in the hands of the ehiefs and the priestsofOro. This inference is based on the faet that the Tahitian mortuary ritual forhigh chiefs involved the drying of the corpse in the sun, on aplaiform, followedby wrapping (binding) and decoration with red feathers. This was to kecp themighty dead under control, safely dead, and ineapable, because of the massivebindings of bark-cloth enveloping them, from harming the living. Such trcat-ment (wrapping and decoration with feathers) was reserved for eorpses of pow-erful chiefs and for the to '0, so the implication is that the 10 'o, at the conclusionof the pa 'ialua reremony, is being trcated as a 'powerful corpse'. The bindingup of the 10 'o places Oro in the 'patient' position!

FIG.7.6/3. Sorcery ti 'i v.ith hollow back for feathcrs/exuviae.Source: The British Museum, London, MMoZ9679

Feathers, needless to say, are !hee.xuviae of birds, and birds have, everywhef';:in Polynesia (as so often e1sewhere) heavenly associations. Red feathers areprobably also assoeiated with blood, another exuvial substance and index oflife, growth, and reproductivity. I Oro was particularly associated with bothfeathers and redness. The primary act of homage to Oro was the presentationto him of red feathers, which would be woven into his sennit wrapping,attached to it by cords, or simply wrapped inside the covering.

, In Samoa the red feathers woven into the most prestigious fine mats forroing chiefly dowries wereassociated with the hymeneal blood of the virgin taupoll, the predestined reproducer 01' chiefly power.

• Related tO t!lis idea of'binding' the god is the praetice, which was common in antiquity, ofresrrain-ing the images of the gods in tempies with chains and manades, to prevent them from escaping andtransferrtrig their pl'Otection to rival eities (cf. Freedberg 1989: 74-5).

The cult of the gods in the form of to 'o and the accoml'anying featherexchanges can be seen as a kind of vastly magnified exercise in volt sorcery.Ostensibly,Úro rules the universe, but in l'ractice his intervention in humanaffaü's is controlled by the chiefs and l'riests, via indexes which are his l'arts,his exuviae, and the binding of the l'rimary index of his person, the to 'o. It wiUbe recalled that the ~ncient Greek sorcerers wrote on the Iittle lead plates rep-resenting their victims, tht words 'I bind, I bind'. In the cult of the til 'o we findthis 'binding' of the god literally enacted, bur the iml'lications are the same.The prototype of an index is bound to the index by resemblance, and is thussubjected to controI. The io 'o were themselves tightly bound, and their powerdiffused into these bindings, especial1y the feathers, which then became thecurrency of l'olitical controI.

This idea can l'erhaps be given a more general statement, so as to apply toali religious art. The great monuments that wc~have erected to God, the greatbasilicas and cathedrals, are indexes from which we abduct God's agency overthe world, and over his mortal subjects, who have striven and laboured toplease him, and have left these massive shel1s (01' skins) in their wake, withinwhich the faithful gather to worship the ultimate author of al1 this magnific-ence. Such, at any rate, is the orthodox view of religious magnificence, whichrel'resents basilicas and cathedrals as 'offerings' to God, who is al1-powerful.However, it must also be recognized, first of ali, that God is not really powerfulat ali unless his power is apparent in this-worldly indexes (behavioural ones,01', in the present case, material ones). The basic Miltonic l'aradox that God isat risk from his own creation, simply by having distributed himself in manifoldforms (including Satan), applies to such material indexes ofGod's greatness inthe fol1owingway.li.llmanity has a Iien on God because his objectification is intheir hands. Even if God is the ultimate author of his resemblance in the tormõfmagnificent structures and works of art, it remains the case that, at a crit-ical point in the sequence of causes, instruments, and results, human agencyis essential. Since, in this world, God's presence is inherent in these works oflÍúman agency, he is bound to human purposes, the this-worldly prosperityand other-worldly salvation of his ostensible servants, rather than to purposesentirely his own. His agency is enmeshed in ours, by virtue of our capacity tomake (and be) his simulacrum. With respect to our god, we are in just the sameposition as the Maori tohunga with respect to the hau of the forest.

Of course, the filct that ~e have tral'ped God inside his Iikenesses does notmake al1 religious activity sorcery, as such. The homage l'aid to God in thebasilica is not destructive ar malignant, but it does make him the 'l'assive agent'of essentially human designs, just as the homage l'aid to the to 'o made Oro thepàssive agent of the chiefs and high priests. 1 have no doubt that Christian the-ologians would have no difficulty in refuting such an iml'utation in a mannerconvincing, at least, to themselves, but the logic of the situation seems to me

mescal'able. The papal tide 'Pontifex Maximus' (JVLlVS 11 PONTIFEX MAXIMVSblazoned over the apostle of si Peter's in Rome) attributes to the Pope thepower to build bridges between earth and heaven. The analogy between thepapal 'bridge' between earth and heaven, and the Tahitian to 'o-the stafl' orprol' ~~ich keel's heaven and earth al'art (but aÍso in communication)-seemstoo remarkable to pass without comment. And in St Peter's one can certainlydetect the c1ear implication that God is the exchimge-l'artner of his moreiml'ortant subjects such as Julius 11if not the mass of his worshippers of lowestate. St Peter's is the bridge; but the point is that the making of the bridgehas beenattributcd unambiguously to the Pol'e, Julius 11, his l'redecessorsand succ~ssors. He, in exchange terms, is the primary donor, the holder of thekitoum, the 'unencumbered valuable' (Munn; see below, Sect. 9.3) which issent out to find its match, the valuable which can be measured against it andreturned for it. St Peter's, as a gift-object and an index of human agency, eli-cits a responsive counter-gift, which, paradoxically, is St Peter's itself, investedwith divine power now available to mankind, like the feathers which, given toOro, return from Oro as his embodiments.

However, let us not delay too long over such analogies, which may beregarded as unconvincing, 01' even offensive. 1 now want to turn to the wor-shipofanthropomorphic images, among which neither the to 'o nor St Peter'smay be counted, at least, not àt first glance.3 The literature of idol-worshil'is, in the main, profoundly itnsYtnpathetic to this l'ractice; it is almost as iflearning to read and write disqualifies one from engaging in this practice withany enthusiasm. This rather sul'ports the notion that there is a 'great divide'b~tween the essentially non-sensuous mode of Iiterate thinking and the sensu-ous, partíéipatory mode of l're-literate thinking. However, there has been areaction agairist the theory of the 'great divide' in recent years (Parry 1985) andrhere certainly exist numerous literate image-worshipl'ers in the world today,and andent, Iiterate, éivil1zations whose religious l'ractices centre on the pay-ing ofhomage to images, such as Hinduism. The idea that only the uneducatedor 'primitive' worshil' idols of sione, wood, and metal fashioned to resemblethe human form, is a consequence of the convergence between !\nti-imagistic !forms of religiosity (such as Judaism, Islam, and certain forms ofChristian sec- \fãiiaiiismarid Pl'otestantism) and the rise of a more generalized religious scep-ticism, which has ancient antecedents. Indeed, wherever religion exists, it isprobable that s~epti~~sll1also exists, whether 01' not this is exl'ressed in l'ublic.Certainly, antnropologists have encountered innumerable examl'les of scepti-cism concerning the efficacy of rites among the illiterate and uneducated, sothis cannot be attributed to Iiteracy 01' the rise of 'science' alone.

, But ef. Wittkower 00 anthropomorphie elements in Renaissanee theories of arehiteetural propor-tion, Aekerman on Bernini's piazza, ele.

7.7. Darshan: Witnessing as AgencyQuestions of the cultural foundations of belief and scepticism need not detainus, but it is still enormously difficult for Westerners and non-believers toempathize with idol-worshippers because of the bombardment with anti-idolatrous propaganda which we have experienced from the very moment webecame conscious of such things. Perhaps the most insight-provoking accountsof image-worship belong to the literature of Hinduism, because followers ofthis religion are among the least self-conscious'in showing devotion to images.A convenient int~óduction to the copious literature on image-worship in Indiahas been provided by Diana Eck (1985)' Worshipping images is obtainingdarshan from the god, a particular type of blessing conveyed through the eyes.Darshan is something given by the god, a mode of the god's agency in the world,and the worshipper is a patient (the Recipient··P, in terms of our scheme).Living human beings can give darshan, as well as gods; a guru gives darshanwhen he or she makes an appearance before a gathering of disciples (see Babb1987), and the same is true of ao important politician appearing before anassembly of supporters, who have cometo see, as much as to hear, their leader.Darshan is a gift or an offering, made by the superior to the inferior, and itcõnsists of the 'gift of appearance' imagined as a material transfer of some

; blessing.It seems to me that darshan is essentially simila.rto the other made in which

divine blessing/personhood is distributed inindia: the distribution of sacredfood, prasad, which is consumed bythe god's devotees (of course, there areChristian parallels to this as well). Prasad is often conceptualized as the 'foodleavings' (jutha) of the god, in a manner absolutely analogous to the exuviaeused in sorcery. According to Eck, the conceptualizatlon of darshan is closelyallied to the role allotted in the Hindu tradition to the eye as an organ of inter-personal transactions .

.Darshan, considered as a mode of divine agency is thus intimately connectedto the concept of the, evil eye.Divine idols, religious gurus, and politicians ofrenown transfer blessings via the steady and penetrating gaze with which theyirradiatethe assembly over which they preside. This is, so to speak, the posit-ive 'white' aspect of evil-eye sorcery, which more imperfect beings transfer viatheir mean, envious, and ill-intentioned looks. To place oneselfbefore the idolof the god, therefore, is to lay oneself open to the divine gaze and to internalizethe divine image.

This is to examine the question from only one angle though; even in India,it takes an act of will on the part of the worshipper to worship the god, andthe worshipper is also an agent with respect to the one being worshipped. Thereception (as a patient) of darshan from the god is contingent on the transitiveaction of 'taking' (darshan lena) initiated by the recipient. Eck cites SteHaKrarnrisch's account of 'seeing' as transitive form of agency:

Seeing... is a goingforth ofthe sight towardsthe object. Sight touchesit and acquiresits formoTouch is the ultimateconnectionby whichthe visibleyieldsto beinggrasped.While the eye touchl.'Sthe object the vitality that pulsates in it is communicated ...(Kramrisch 1976: lJ6)

These remarks of Kramrisch are informed by her knowledge of Sanskrit philo-sophical writings. Ancient Indian philosophers held views similar to thase ofinePlatonists, that sight was an 'extromissive' sense, the eye sending out invis-i51ebeams or rays through the air,whichtouched the objects of sight at theirsurfaces. One philosopher, Caraka, argued that indeed we only have one sense,ihe sense of touch, of which sight, hearing, etc. are just more subtle forms (Sinha1934}:Other philosophers disagreed, but the consensus was none the less thatseeing was, like touching, a form of contact. Epicurus' theory (see aboveJ, bycontrast, is iintromissive'; the tidola emanate fTOmthe object and enter the eye.But Lucretiús shows that seeing was, even so, equated with touching by theEpicureans. The (gossamer, but physical) idols were touched at the surface ofthe eye, not at their own surfaces, via the eye-beams. I do not think that theEpicurean intromissive theory of vision by 'idols' had an Indian counterpart, 50

one could not attribute the prevalence of image-worship to its influence. But thealternative extromissi~e t~~ory is, if anything, even more explanatory, becauseit forges a díiect link between 'seeing' (darshan) and other types of physicalinteráction with the image, such as touching, anointing, and so on. These taetileformsôf homage are very important elements in Hindu image-worship.

Certain Indian philosophers (Sinha 1934) made the analogy between seeingand the use of a stick by the blind, in order to ascertaÍll the shape of externalõbTects.This materialistic conception of seeing is reflected in Kramrisch's5tatement about darshan; seeing creates a physical bridge between one beingand another. Hirn's basic insight that in relation to 'images' there really isno distinction between 'similarity' and 'contact' is fully brought into the openhere, without the need to invoke the Epicurean paraUe!.

Darshan is thus very much of a two-way affair. The gaze directed by thegod towards the worshipper confers his blessing; conversely, the worshipperreaches out and touches the god. The result is union with the god, a mergingof consciousnesses according to the devotionalist interpretation. This bringsoa-êki:he issue of reciprocity and intersubjectivity in the relationship betweenthe image (the index) and the recipient. It is c1earlygermane to the general the-sis argued in these chapters that intersubjectivity between persons and indexes,particularly indexes which, like images of the gods, are human in form, should,?-e.possible. It cannot be denied that, from the point of view of the devoteesworshipping the image of the god, the image of the god is a manifestation of asocial Other, and that the god/devotee relationship is a social one, absolutelycomparable to the relationship between the devotee and another human per-son. However, it is too easy just to accept this as an ethnographic, descriptive,

fact, without a deeper questioning of the cOI~nitivebasis of this relationship.For ali that the devotee asserts, and truly bdieves, that a union of minds i~achieved between mortal devotee and immonal divinity, the devotee none theless also !ives in a world of ordinary objects, mere things devoid of impute(:subjectivity, in which the distinction betwcen human beings-possessed ofhuman-being-Iike consdousness and agency-and inert 'things' is readilydrawn. The devotee does know that the image ofthe god is only an image, not~ade of flesh and blood; and if, perchance, the image moves, or speaks, orseems to drink milk, that is a miracle, a remarkable occurrence because soupexpected. Devotion is enhanced by such rnanifestations, but it is admittedthat the gods only vouchsafe miracles when faith in them has reached a lowebb, and needs to be buttressed by extraordinary happenings. True devotioni:;àttainable, ideally, without miracles of this or any otherkind.

The animacy and imputed subjectivity of the idol is not attained except bysurmounting the stark difference between an inert image and a living being.How does this happen?

So far as the Hindu material is concerned, the key to theprocess of anima-tion seems, initially at least, to depend on the lô'giê of looking and being seen.Imagistic devotion is a visual act (as opposed to prayer, etc.) and it is accom-plished entirely by looking. Specifically, it is accomplished by looking into theeyes of the god; union comes from eye contact, not the study of al1 the otherdetails the image may show, which indicate the identity and attributes of thegod and which add to the general effect without being devotionally essential.The eyes of the god, which gaze at the devotee, mirror the action of the devo-tee, who gazes at the god. Sometimes (as in Jain temples) the eyes ofimages areset with little mirrors, so that the devotee cansee himself or herself reflectedin the image's ~yein the act oflooking. Even in the absence of actual mirrors,the image, so far as its ocular activity goes, reflects the action of the devotee(Fig.7-7/I).

Animacy takes its origin from this ocular exchange, because, even if one doesnot take a mystical attitude towards images, one is none the less entitled toapply action verbs like 'Iook' (or 'smile', 'gesticulate', etc.) to them. A perfectsceptic can say, in fact is obliged to say, that an idol 'Iooks' in a particular direc-tion; the remark would pass unnoticed because everybody accepts that the cri-terion for idols 'Iooking' is that their eyes should be open and pointed in aparticular direction. The question is, what do idols see when they look? Whatthe devotee sees is the idollooking at him or her, performing an act of ióoking,mirroring his or her own. Ir is not mysticism on the devotee's part whlchresults in the practical inference that the image 'sees' the devotee, because weonly ever know what other persons are seeing by knowing what they are look-ing at. The sceptic would say 'the idol is blind-it cannot see anything', buteven so, to be blind is to be unable to see what one looks at, which hardlybanishes the residual animacy of images, since a disability implies a potential

FIG.7.7/1. Santoshi Ma. Colou r lithograph, B. G. Sharma. 50/l",e: Sharma Picture Publications,Dombay, L1960

• I • ' '. ~" ,, - - ~ _~ __ ~_ - - • I

Reclplent-A ~ Index-PDevotee sees ~ Idol

~Index-A -... Reclplent-P

Idol sees Devotee

~Reclplent-A Index-P

Devotee -... Ido Isees~ FIG. 7.7/2. Ocular exchange

as lhe rnediurn forintersubjeclivities

ability. And the devotee need not draw this inference, in fact, cannot, bccausethe situation is defined in terms of the devotee's own agency and result; thedevotee looks and sees. The image-as-mirror is doing what the devotee isdoing, therefore, the image also looks and sees. One does not even logicallyh;v~ to impute 'Üfe' to an image to assert that an image can see; after ali, peo-pie often speak of cameras as 'seeing' things, without implying that camerashave life.

But the inference that if idols can look, they can see, is not drawn explicitly;if it were, it might be open to sceptical objections. It is protected from judge-ment by a further consequence of the mirror-effeet. This is the logical regres-sion set uphyseeingandbeing seen. Eye-contact, mutuallooking, is a basiemechanlsm for'intersubjectivlty 6ecause to look into another's eycs is not just

Y to see the other, but to see the other seeing you (I see you see me see you seeme etc.). Eye-contact prompts self-awareness ofhow one appears to the other,at which point one sees oneself 'from the outside' as if one were, onese1f, anobject (01' an idol). Eye-contact seems to give direct access to other mindsbecause the subject sees herself as an object, from the point of view of the otheras a subject. Eye-contact is the basic modality of 'second-order intentionality',awareness of the other (person) as an intentional subject. Thus, in image-worship, the devotee does not just see the idoI, but sees herself (as an object)being seen by the idol (as a subject). The idol's 'seeing' is built into the devotee'sown self-awareness at one remove as the object w hich is seen by the ido!. Shesees herself as the idol sees heI', kneeling before it, gazing upwards. In that shecan see herself seeing the idol, the idol must see heI', because when she sees her-self seeing the idol (from heI' point of view, a datllm of immediate experience)the idol is seen by heI' as seeing heI'. The 'idol seeing heI" is a nested componentof 'heI' seeing herseIf seeing the idol'. The net result of the regression wherebydevotee's and idol's perspectives become logicallyinterdigitated with one anotherin this way is a kind of optical oscillation in which idol's and devotee's per-spectives shift back and forth with such rapidity that interpersonal boundariesare effaced and 'union' is achieved. This mutllal encompassment via eye-contact is represented diagrammatically in Fig. 7.7/2. I shall return to the

significant role ofthe 'eyes' in idol-worship, below (Sect. 7.12) in connectionwith the consecration procedures for images.

7.8. Animism and AnthropomorphismThe worship of images, 01' aniconic indexes of divine presence, such as stones,springs, and trees, has preoccupied anthropologists from the beginning. Tylor,in the spirit of his times, sought an explanation in a purported difference of'mentalities' betwcen primitive and civiIized people; primitive people wereanimists, whereas we are noto TyIor imagined that the belief in spirits, and 'thesupernatural' generally, was implanted in the human mind through the experi-ence of dreams, and arose from conceptual confusion, but this was merely aself-serving supposition on his parto In effect, the theory of animism is merelya classificatory deviee, which, like the concept of ritual, serves to separate beha-viour that we think we can understand and sympathize with, from behaviourwhich seems to us superstitious and perverse.

However, the Tylorian concept of animism can be made into a more ser-viceable analytical tool if it is abstracted from the essentially pcjorative contextof Victorian positivistic thought. Guthrie, in a recent study of the founda-tions af religious belief, has argued that 'anthropomorphism'-the tendencytõ impute human attributes such as will, agency, and responsiveness to sup-posedly 'inanimate' entities-is an abiding feature ofhuman cognition (Guthrie1993). His cognitivc argumentis that, strategically, it is aIways safeI' to imputethe highest degree of organization possible (such as animacy) to any givenobject of experience. It is better, he says, to presume that a boulder is a bear(and be wrong) than to presume that a bear is a boulder (and be wrong). Hisargument carries weight, and he certainly has no difficulty whatsoever in amas-sing copious examples of anthropomorphism, not just in religious contexts, butin everyday perception and cognition, in the arts, and even in the sciences. Quiterightly, Guthrie emphasizes that anthropomorphism is not a phenomenonrestricted to children, so-called 'animists', 01' even to adult religious believers.As a non-religious adult, and a participant in an advanced technological civil-ization, I recognize that I consistently engage in anthropomorphic thinking,as my earlier remarks about cars will have made apparent. The trouble withGuthrie's argument is not that it lacks an empirical basis-far from it-but thatto say that one attributes 'animacy' 01' 'anthropomorphism' to something doesnot explain what a thing must be 01' do to count as 'animate' 01' 'anthropomorphic'.

It is not animism, anthropomorphism, 01' anything like it, to attribute 'life'to a trce, which adults in our society agree is a living thing-though childrenunder 5 01' so may disagree (Carey 1985). On the other hand, it certainly isanimism to attribute the capacity to 'hear' prayers to a stone idol, but thatdoes not necessarily imply the belief that the stone 01' idol is 'alive' in a' bio-!ogica! sense. This is abundantly evident from the furore which results when

it is reported, as sometimes happens, thar idols actually do 'come to !ife' inthe sense of'manifesting biological activity'. When particular statues bleed, orperspire, or move about, these are 'miraeles'. But such happenings would notbe miraeles if the expectation was that ali idols should behave in this wayjin fact, they are generally expected not to. Idols may be animate without, mother words, being endowed with animal Jife or activity. In addition, thtreare automata, real or imaginary ones, which do really move, speak, and perforrnvarious human-like actions, but these are remarkable, not because they arealive, but because they are not alive, while maintaining the appearance ofbei:lgso. !t follows that 'ritual' animacy and the possession of 'life' in a biologicalsense are far from being the same thing.- Cognitive psychologists believe that the distinction between Ii'ling things(biological organisms) and non-living things is one to which we are innatdysensitive, and this is a finding which I do not need to contesto I have no doubtthat worshippers, who address their prayers to stones, are perfectly capable ofthe category distinction between animais and plants (living things) and nOI1-living things such as stones. The distinctiOI1between animacy and inanimacythat we require here cross-cuts the distinction between the living and non-living as 'natural kinds' (Boyer 1996: 86). The god who, at one moment, IS

manifested in a non-Iiving thing, such as a stone or a statue, may be manifestin a living thing also, such as a possessed shaman, or a sacred goat or monkey.Tfie worshippers, whose god appears in these contrasted forms, are perfectly

. cognizant of the difference between them. The imputation of 'animacy' to non-living things cannot, as Guthrie seems to sllggest, rest on people making cal-egory mistakes about whether inanimate objects such as boulders are reallvbiologically living things, such as bears. .

made of a~:counting for 'intentionality' even among ordinary, living, breathing,human beings.

The question as to the manner in which representational indexes-idols--can be apprehended as social others, as repositories of agency and sensibility,seems to raise the question of 'apparently irrational' beliefs and practices. It issurely irrational, or at least strange, to speak to, offer food to, dress and bathea mere piece of sculpture, rather than a living breathing human being. And soit is: those who do these things are just as aware of the 'strangeness' of theirbehaviour as we are, but they also hold, which we do not, that the cult af theidol is religiously efficacious, and wiII result in beneficial consequences forthemselves and the masters they serve in their capacity as priests. Ir is not thatthe priests cannot distinguish between stocks and stones and persons, rather,they hold that in certain contexts stocks and stones possess unusual, occult,properties; of which the religiously uninstructed would remain ignorant, andthe instructed but sceptical, wrong-headedly incredulous. According to Boyer(1996), religiolls ideas, such as the 'intentional psychology' of fetishes andidols, survive and prosper as components of cultural systems just because theyare odd and counter-intuitive.

The question that we need to consider is the nature of the unusual occultcapacities whkh idols possess, according to believers. What we need to knowis how idol-worshippers square the cirele between 'what they know'-and whatwe know they know-about stocks and stones, and what they know about per-sons and their capacities as intentional agents. They cannot confuse the two,but it remains possible that persons have attributes which can be also possessedby stocks and stones without prejudice to their categorical difference fram per-sons. That is 10 say 'social agents' can be drawn from categories which are asdifferent as chalk and cheese (in fact, rather more different) because 'socialagency' is not defined in terms of'basic' biological attributes (such as inanimatething vs. incarnate person) but is relational-it does not matter, in ascribing'social agent' status, what a thing (or a person) tis' in itself; what matters iswhere it stands in a network of social relations. Ali that may be necessary forstõCksand stones to become 'social agents' in the sense that we require, is thatthere should be actua! human persons/agents 'in the neighbourhood' of theseinert objects, not that they should be biologicalIy human persons themselves.

This is not as bizarre a elaim as it sounds. We certainly do not have to pos-tulate a particular 'mentality' (primitive, uncritical, gullible, etc.) to accountfor idolatry; the worship of images is compatible with an extreme degree ofphilosophical and critical acumen, as the example of textual Hinduism amplydemonstrates, besides numerous treatises on 'theurgy' (the creation of gods)from elassical antiquity (e.g. Proelus). Rather than approach the problem froman initial assumption of the essential stupidity of idol-worshippers, we shouldremind ourselves ofthe difficulties which assail (very elever) philosophers, whcnthey seek to account for the 'agency', not of stocks and stones, but of human

It seems that we must appeal to a different attribute from the possession ofbiologicallife to define 'animacy'. Should we say that the object is animatenot because we attribute biological life to it, but subjectivity/intentionality,which is something quite different? As;, Boyer notes (1996: 92), 'the projec-tion of human physical features [onto gods, spirits, etc.] in general results in aprojection, tacit or otherwise, of intentional psychology, but a projection ofintentional psychology does not generally entail a projection of any otherhuman quality'. How can an entity possess 'intentional psychology', without~_~ingbiologicalIy alive? In this case, a worshipper who addresses prayers to astone must believe, somehow, that the stonc in question, though not a livingthing, sees and hears as does the worshipper, thinks and reacts as he does, andmoreover, has the power to plan and execute actions. It seems paradoxical toimagine that an admittedly non-living thing could possess these attributes.But not realIy, when one considers what heavy weather philosophers have

beings themselves. If (Western) philosophers have a hard time pin-pointingexaetly what makes the differenee between human persons engaging in 'aetiofls'and mere things obeying causallaws then we are in a better position to understatldwhy some people appear to be (eontextually) indifferent to this distinetion.

The faet is that it remains a eontroversia! philosophieal problem to distin-guish between 'aetions' (stemming from ageney) and 'happenings' resultingfrom material eausation. What is the basis of the distinetion we feel we mustdraw between a person performing the aetion of raising their arm (intention-ally) and the same physieal movement occurring involuntarily, say, as theresult of some malfunetioning of the autonomous nervous system? In otherwords, there is a sense in whieh human beings are themselves 'stoeks andstones'-only rather twitehy ones-and wh':f\ human beings are asleep, insen-sible, 01' of eourse, dead, the resemblanee beeomes mueh closer. The 'philo-sophy of aetion' is devoted to devising and testing eriteria for justifying thedistinetion that we intuitively make between our eapaeity to behave as pel'-son/agents and our simultaneous eapaeity to be, and behave as, things, 01' mere'ereatures' unendowed with full human ageney.

I do not need to eanvass, still less evaluate, ali the theses whieh differentphilosophers have advaneed in the seareh for a solution to this problem. Mypoint is just that philosophers do deteet a serious problem here; from whieh Ithink that we ean reasonably infer that it is in faet awkward to differentiatebetween what people do 'as persons' and what people do 'as things'. And ifphilosophers (who know perfeetly well that in the relevant senses persons andi:hings are different) eannot always agree about 'just how' they differ, then theinverse proposition also holds; if pressed, we are not really very sure about 'justhow' an idol is no! a person-even though we are perfeetly eertain it isn't,We eannot rely on simpie arguments like this: 'I am a person, I have a beatingheart, a temperature of 98.4 degrees, ete. whereas this statue is stone-eoldand has no heartbeat, ergo, it is not a person'. We eould heat the statue to98.4 degrees, give it a heart, ete. but it still would not qualify because it wouldonly possess, after these modifieations, attributes whieh human beings pos-sess 'as things', not as persons/agents. Yet sueh are the argumellts whieh areadvaneed against idolatry by seepties:

mouse eould equally well have been interpreted as his superhuman ascetieismmanifesting itself-indeed eomparable feats have been attempted by ascetieswhose humanity is not in doubt. Some, for instanee, are reputed to haveindueed birds to nest on their motionless bodies. Any sueh argument can eutboth ways: if idols are not what they pretend to be, 01' are pretended to be, itis not beeausc they are 'things'. Human beings are also 'things'. If an effigypossessed every single thing-attribute of a human being, it eould still be 'justan effigy' and be unworthy of worship; eonversely, an effigy eould possessno identifiable thing-attribute of a human being and be worthy of devotionnevertheless.

Suppose we perform tb.e thought experiment of gradually enriehing an idolwith more and more of the attributes of a 'genuine' living being. Temperature,heartbeat, mobility, the ability to utter words, to play tennis, to ... well, youname it. Does the idol beeome more worship-able as a result? By no means:either, bythis proeedure, the idol beeomes a eommon-or-garden human per-son, whom it would be senseless to worship, 01' the idol remains an idol buttakes on the status of an automaton of extraordinary verisimilitude, worthyof exhibition and admiration in an establishment like Disneyland 01' MadameTussaud's, but not of reverenee ar devotion. The eriticism of idolatry on thegrounds that idols are not 'alive' as human beings are (biologieally) alive, 01'

that idols are not realistie automata, but only statues, misses the point on botheounts. The idol is worshipped beeause it is neither a person, nor a miraeulousmaehine, but a god.

Whatever the attributes possessed by idols whieh render them religiouslyeffieaeious as a loeus for person-to-person eneounters with divinities, theseattributes eannot be eonfirmed 01' disproved by physical tests sueh as the pres-enee of a pulse, respiration, ingestion and e1imination, the ability to move 01'

speak, a natural distaste for miee, and 50 on. None of these attributes figures inphilosophieal attempts to distinguish between person/agents and mere things,maehines, effigies, iIIusions, and so on. Currently, many philosophers agreethat 'ageney' implies the possession of a mind whieh 'inte,.,ris' aetions prior toperforming them. 'Not moving' is an 'aetion' in this sense. 'Shiva (the god)did not move because he intended to stay still' is a perfeetly reasonable inter-pretation for the seene witnessed by Dayananda, on the assumption that Shivawas, as the priests averred, present as a person/agent, in his image. He mighthave had various reasons for refraining from aetion; first (as just mentioned),beeause Shiva is the prototypieal immobile aseetie, and seeondly, beeauseShiva (who ereated the whole world) imbued stone, the material of his visible'body' in this instanee, with the property of absolute rigidity; Shiva wasobserving the 'roles' for stone objeets, sueh as idols, whieh were of his ownmaking, ultimately. Hindu theology, moreover, postulates that gods, sueh asShiva, voluntarily 'saerifiee' their freedom of movcment, imprisoning them-selves in stone idols for the benefit of devotees.

Laia Lajpat Rai describes how the founder [ofthe Arya Samaj, a 'reformed' Hindu sec!,opposed to the use of images] first got his insight imo the wrongness of idolatry. Hewas set, as a lad of fourteen, to watch an image of the god Shiva, in a tem pie at night.He saw a mouse run over the god's body and the god remain motionless. The shockconvinced him, Lajpat Rai wrote, that 'the image could not be Shiva himself, as wastaught by the priesthood'. (Bevan 1940: 34)

Sueh an argument could have been convineing only to one who had already(no doubt as a result of Christian-Protestant aseendaney in British India)deeided that idol-worship was backward and futile. Shiva's indifferenee to the

This is the greatest grace of the Lord, that being free He becomes bound, being inde-pendent He becomes dependem for ali His service on His devotee ... In other formsthe man belonged to God but behold the supreme sacrifice of Isvara [Vishnu) here theAlmighty becomes the propcrty of the devotee. . He carries Him about, fans Him,feeds Him, plays with Him-yea, the Infinite has hecome finite, that the child soul ma)'grasp, understand and love Him. (Pillai Lokacarya, cited in Eck 1985: 46, diacriticsremoved)

There are no a priori logico-philosophical grounds for rejecting the thesisthat Shiva 'intended to stand stilI, and thus stood stilJ, when the mouse J'allover him' (and was an agent in so doing). Because there are no 'material tests'for the possession, or non-possession, of agency, there is nothing to preventus from asserting, if we wish to, that the behaviour of a statue (standingstilI) occurs because the statue has a mind, intends to stand still, and does as aconsequence of this prior intention stand still.

not immediately square with psychological investigations of the way in whichchildren and ordinary people (not philosophers) seem to approach the sameproblem. It secms that ordinary human beings are 'natural dualists', inclined,more 01' less from day one, to believe in some kind of 'ghost in the machine'and to attribute the behaviour of social others to the mental representations theseothers have 'in {heir heads '. Behaviour is caused by factors which well up fromwithin the person, thoughts, wishes, intentions, etc. Minds are .hidden aw~yinside people, rather than being manifested in bet',Veen them, m the pubhcspace in which interaction takes place, as the externahst theo.r~ seems to be ~ay-ing. The 'internalist' theory of mind, according to the cogmtlve p.sychol?glsts,is a 'module'-·-a kind of theory (or principIe of interpretation) wlth whlch weare born along with the principIe that there is a basic distinction betweenliving things and non-living things. When Boyer speaks of'intentional psycho-logy'he is referring to this (putatively innate) module. .

~any philosophers believe that the notion that genu~ne 'persons' are bel~~swhose behaviour is caused by the mental representatlons that they have mtheir heads' is not just a truth of common sense, but ~sperfectly defensiblephilosophically, if formulated with due care (e.g. Fodor 1994)· !'I0,:ev~r, weaonot need to consider the arguments for and against the 'mentahst' (I.e. mter-nalist) position in the philosophy of mind in any d~tail. ~Il t~at we ~eed to beaware of is that these two routes towards 'agency -attnbutlons eXlst. Let usconsider furt:her some of the differences between them. For a start, they eachbegin with a rather different problem-definition. The externalis.t theory.is ~otreally about the 'psyche' or 'consciousness'; it is an.ac~ount o: mtersu.ble~t~v-ity rather than subjectivity, and it explains how It IS ~hat mtersubleCtl~ltyis quite possible, even in the absence of some teIe~athlc. me~ns of entennganother person's skulI and having his thoughts, feehng hls pams, and so on.Because the externalist theory is about intersubjectivity, it is popular with soci-

, ologists, many of whom are llluch more behaviouristic in t~eir thi~king th~nthey realize or acknowledge. To cite a case in point, the I~admg s~clal theonstoftoday, Pierre Ilourdit;u, acknowle?ges a debt to b~th ;V~ttgenstem and to the(subtle) behaviourist learning-theonst, Hull. Bourdle~ s mval.uable ~oncept ofthe 'habitus'--··the sedimented residue of past social mteractlOn whlch struc-tures ongoing interaction-is not a transcription of common-s~nse ~ental~smor 'folk psychology', but is prcciseIy a notion of mind externahzed m routme,practices, thal is, the prevailing 'form of \ife'. Sociologists h~ve to ~ 'exter-nalists' becanse culture and social institutions are external, mteractlve, pro-cessual historical realities not states of mind. Sociologists cannot be 'pure'" . .mentalists because, apart from anything eIse, they are concerned wlth act~onsin the light of their consequences, and we are alI tuo welI aware that our actlonsrarely if ever have precisely the conseq~ences w~ hoped ?r e~pecte? them tohave. So a theory which only relates actlons to (mner, pnor) mtentlons, evenif adequate psychologicalIy, is sociologicalIy inadequate.

7 ·9· Externa! and Interna! Conceptions o/ Agency

How, in practice, do we attribute 'agency', 'intentional psychology'-the pos-session of a mind, consciousness, etc. to 'social others'? If we knew the answerto this question we might be in a better position to define preciseIy what sub-set ofhuman-like attributes idols as 'social others' are believed to possess, giventhat they are probably not (ou Boyer's argument) believed to possess alI ofthese attributes. I think it is fair to say that our attribution of 'intentional psy-chology' to anything (a person, an animal, a computer, a cal', or a stone idol)has two aspects to it, which at first glance seem to be rather distinct. The firstof these one could call the 'external aspect' or the 'practical' aspect of agency-~ttribution. According to Wittgenstein, and a great many other subsequentphilosophers, the possession of a mind is something we attribute to others,provisionalIy, on the basis of our intuition that their behaviour (e.g. theirlinguistic behaviour) follows some 'rule' which, in principIe, we may recon-struct (Winch 1958). If I can get along with lhe other in the give-and-take ofinteraction, if our practical efforts to deal with one another work out, then theõther is a producer of intelIigible (meaningful) behaviour, and hence has amind, intentions, volitions, etc. I cannot really telI, from the outside, whetherthe 'other' is a zombie or an automaton, who/which mimics the behaviour ofan ordinary human being but does not have any of the 'inner experiences' wehabitualIy associate with this behaviour. But this does not matter becauscthe whole panoply of 'Illind' is not a series of inner, private experiences at alI,but is out there, in the public domain, as language, practices, routines, rulesof thegime, etc.; that is, 'forms of life'. CalI this the 'externalist' theory ofagency-attribution.

However compelIing in the hands of behaviouristically inclined philoso-phers, the 'externalist' theory of agency has a weakness, nameIy, that it doe:;

· Sinc~ t,his is a treatise on the sociology of art, there are good reasons why I,m particular, should pay attention to the externalist theory of agency. Thcsimplest solution to the problem of idolatry is an 'externa!' one on the foIlow-ing lines: idols are 'social others' to the extent that, and because, they obeythe social rules laid down for idols as co-present others (gods) in idol-form.Thus, according to Eck (1985: 48), a Hindu deity in idol-form is esscntial1y an~onoured 'guest' to whom the devotee pays homage in the forro of nurture(offering food, fanning the flies away, etc,). The appropriate behaviour for'guests' of extremely high status, is, in fact, to do more or less exactly whatidols do: accept what is offered with imperturbable dignity and impassivity.Idols 'produce inteIligible behaviour' which conforms to certain expectations.

Of course, idols do not apparently 'do' anything; they generaIly just standthere, being immobile. This seems Iike an odd form of'intelligible' behaviour,but it is not; the J;.ifeGuards outside WhitehaIl barracks produce exactly (histype of 'intel1igible' immobility and apparenl insensibility as a behaviour, andthey are quintessentiaIly 'social others' while they do so. When evil-mindedtourists poke umbrel1as at their horses, they do not suddenly behave 'out 01'character' and curse their tormentors: they preserve their icy indifference.They are playing the game, and it is a game we can readily participa te in. Theyare like the Sepik warriors whom Harrison describes as aspiring, in totemicritual performances, to emulate the impervÍous spirituality of totemic sacra,the ceremonial effigies of spirits which line the interior of their long-housc(Harrison 1983: II8).

confounding traitors and plotters, or keeping the sun alight, or enjoying t~~m-selves in heaven, ete. And these actions (except perhaps the last) have vlslbleconsequences 'elsewhere', though not in any change in the bodily demeanouróf the idol per se. '" .

Secondly, ali idol who does not respond actlvely (by movmg or speakmg) ~snone the less 'active' as a patient with respect to the agency of others. And thlsmay be enough. Children's play with ~olls can serve to iIIu.strate this .ty~ ofpassive agency. Ool1s, even those which open and c.losethe~r e~es, emlt. crtes,oreven wet themselves never produce any behavlour whlch IS not dlrectlyunder the control of th~ nurture-providing play-mother. The playing childknows this perfectly welI, but that does not prevent them having the Iivelie~tsensation that the doll is an alter ego and a significant social other. 0011 play IS

so total1y satisfying just because of this passivitYi the dolI does just whateverthe child wants' submits to undressing and getting dressed again, sleeping,waking, and eati~g, and even, ifnecessary, being 'naughty.' whe~ the ch.ildis inthe mood to inflict a smacking. The dol1's 'thoughts' and mner hfe (whlch cer-tainly are attributed to it while play is in pro~ress). are a reflex .of ~he.child'sown thoughts, which include the doll as a passlve bemg wh~se thmkl?g IS done'for it' just as dressing and undressing have to be don~ for It as wcll. fhe.play-ing child thinks for her dol1 as welI as doing everyt~mg else. 'fe could tnt~r-pret the thought-processes of idols along the same lmes, .that IS,.assomethtngthat devotees do for idols, which can none the less be attrtbuted (m context) tothem. Because idols (Iike dolls) are wholly 'passive' others, they exhibit 'pas-sive agency', the kind of agency attributable to soci~1others who or which, bydefinition, are only the target of agency, never the mdependent source. (Theargument of this paragraph relates to the analysis of Hindu darshan, above,Sect·7·7·)

However, I agree that neither of these responses quite measures up to theobjection that idols, because of their behavioural ineffectuality, ca~no~be con-sraered 'agems' in the fulI sense. There does seem to be a baslc dlfferencebetween the idol-Iike Life Guard whose mind seethes with unuttered curses,and the genuine idol whose dignified gaze betokens, for us, no such inner life.Whoever imagines that the idol is conscious, thinking, intentional, etc. isattributing 'mental states' to the idol which have implications, not just ~orthe external relations between the idol and the devotee (and the form of hfe1t1 which thev co-parti('1pate), ~ut for the 'inner structure' of the idol, thatis, that it has'something inside it'which thinks' or 'with which it thinks'. Theidof may not be biologically a 'living thing' but, if ir has 'intentional psycho-logy' attributcd to it, then it has something Iike a spirit, a soul, an ego, lodgedwithin it.

This is certainly true, ethnographicalIy and psychologicalIy, because of theinnateness of the 'theory-of-mind module' which attributes intentionality topersons (and things as well, under certain circumstances) as a component of

This 'Dionysian' aggrandisement ofthe self, reserved for a select few, is viewed not asa celebration of subjectivity but as the reverse: the depersonalisation of the actors intohuman equivalents of ritual objects, like the masks, statuary and other sacra whichfigcre in the men's cult as embodiments of ritual potency. Men contextually suspendtheir 'normal' social identities but gain power, in the form of the impact which highly-charged symbolic objects have on the subjectivitics of others.

tIowever, I agree that such an externalist interpretation of agency-attributionto idols seems too simple by half. Although the Life Guards may preserve theirstony demeanours when their horses get poked, none the less, we know, andthey know we know, that they are silently thinking 'bugger off' or words tothat effect. And that is why their self-contral is so remarkable and admir-

i able. !tis all very well to say that an idol which stands immobile is producingintelligible idol-behaviour according to the accepted social rules, but unlessthere is something going on inside the idol which corresponds to i:he 'innerlife, unspoken thoughts, etc.' which we attribute to Life Guards, or Sepikwarriors, then does this behaviour 'count'? There are two answers to this kindof objection. First of all, although idols may f10tproduce much visible beha-viour, they may none the less be very 'active" invisibly, that is, most of their

, actions take place 'off-stage' as it were. They can be making the crops grow,

Fim, the only psychologythat could possiblysucceed in explaining the complexiticsof human activitymust posit internal representations.This premise has been deemedobviousby just abóut everyoneexcept the radical behaviorists... Descartes doubtedalmost everything but this. For the British Empiricists, the internal representation,;were called ideas, sensations, impressions;more recently psychologistshave talked (Ir

hypotheses,maps, schemas, images,propositions,engrams, neural signals,even hol<:r'gramsand whole innate theories. 50 the first premise is quite invulnerable,or at anyrate it has an impressivemandate ... But, second, nothing is intrinsicallya representa··tion of anything;something is a representation onlyfor or to someone; any repres··entation or system of representations thus r~quircs at least one user or interprettl oft.herepresentation who is external to it. Any such interpreter must have avariety ofpsychologicalor intentional traits ... it must be capableof a varietyof comprehension,and must havebeliefsand goals(so it can use the representationto inform itselfand thusassist it in reaching its goals).5uch an interpreter is then a sort of homunculus.

Therefore, psychology without homunculi is impossible. But psychology withhomunculi is doomed to circularity or infinite regress, so psychologyis impossibie'.(Dennett 1979: 119-22)

Dennett argues that this problem is surmountable, not by getting rid 01'homunculi, but by having lots and lots ofthem; 'stupid' single-task homunculidoing low-level tasks anel relaying the results to more intelligent homuneulidoing higher-level processing tasks. There is not any one 'mind' but a pande-monium of homunculi generating representations and selecting among thosethat have been generated those which are of use in fulfilling the organism'sneeds. My purpose in citing Dennett on homunculi, however, is not to intro-duce his bril1iant, artificial-intelligence-based theory of mind, which seeks tobreak the deadlock between the various compelling reasons to accept somemodified form of behaviourism, and the need to explain consciousness as wcactually experience it as psyehological subjects. Dennettt, like alI modernphilosophers of mind, is writing about real human beings, not idols. The poinlthat interests me is that Dennett is suggesting that in so far as we eonceive

of human beings as intentional agents beeause they generate and respond tomental representations, then we are obliged to 'split' them (internaIly) into

\ two' the one who 'has' the representations (pereeptions, ideas, etc.) and the one\. whJ'interprets them (see Dennett, ibid., ch. 5). What Iderive from this is the

cognitive naturalness of the idea of the .'!lind (or soul, spirit, ete,) a~a homun--- culus' that is, like a person but contained within aperson. That lS to say, a

predi~table eonsequence of our (possibly innate) propensity to attribute 'intcn-tional psychology' to humans, animais, etc. is ~ttributing a ho,!,un~ulus-/i~eformto this 'interpreter' lodged within the other, when the other ISbemg attnbutedWlth'an intentional psychology. That is to say, ifwe are to attempt!o 'depict'the physical realization of the other's possession of an intentional psychology,the natural way to do this is to make a duplicate of the other in homuncularform (a representation of the 'inner person who interprets the other's repres-e'ntations) and lodge thaJhomunculus inside the other's body.

Let us return to thefdol/We have established, I hope, that the idol is accept-able as a social other onthe basi5 of'fitting in' to the role expectations for idolsas a-particular category of social agents, that is, l?rimarily passi~e agent~ ~ragents whose agency is exercised 'ofT-stage'. Practically and physlcally, thls.lspêrfectly manageable; we just have to stipulate what the idol, to co~form to ItSrole expcctations, shalI look like, and manufacture an artefact whlch has thestipulated external characteristics. But what about making an ~dol.whieh~ onthe basis of its actual physical characteristies, motivates the attrIbutIon to It ofan intentional psyehology? How might we do that?

Well we might not want to: aceording to Boyer (1996) the attribution ofintenti~nal psychology (or other oceult attributes) to non-livi~g things is ,apõtent religiol1s idea preeisely because it so markedly contradlets t\\'o baslcassumptions about reality, (i) that living and non-living th~ngs are t~t~llydistinct and that (ii) intentional psyehology can only be attrlbuted to hvmgthings. 'On this theory, there would be a strong (basically innate) cognitive

~preference for religious objects (attributed with inten~ional ~sychol~gy) t~ beaniconic in form' the more blatantly the supposedly ammate rItual oblect fallediõmeasure up ;0 the normal criteria for animacy, the more enthusiasticallybelievers would worship it. However, this prediction is not borne out in prae-tice' where the technieal sysrem of a particular religious community includesteêI~niques for manufacturing iconic or anthropomorphic images, .idol~, ete:,such idols are very often manufactured. Moreover, supposedly amcomc reh-gious objects are often 10cally interpreted in 'iconic' ways. ~oodma~'s well-known jibe against the notion of'realism' in art (1976) certaml~ apphes here,namely, that sipce everything 'rese~bles' eve~ything else in at ,leas~ s?m~respects, everything ean, under some mterpretatIon, be regarded.as ~eplctmganything you like. Consequently an uncarved stone ca? be an. Icomc repres-entation of a god just as well as a minutely carved stone Idol whlch looks muchmore 'realistic' to uS.

what Schutz calls 'the natural attitude'. The problems which assail spontaneousmentalism are not to do with demonstrating its existence, but with pursuing itsimplications. Let us say that 'intentional psychology' (that which is atrributt.'dto idols according to the 'interiiâlist' approach to agency) consists of somethinglikea 'conscious self' as experieneed in the 'first person singular'. This is 'the'Mtnd's "I'" (Hofstadter and Dennett 1982). The trouble with this mysteriousT ,isnot that anybody truly disbelieves in it, but that nothing in the world, nophysically identifiable thing, realIy seems to correspond to it. The externalisttheory of mind does not seem to giveone any real reason for believing in thisentity in which we alI do, none the less, believe: and on the basis of such abelief, interpret and predict the social behaviour of others. The philosophicalproblem ofmentalism (ofintentionali5m gencral1y) is suceinctly summed up byDennett in the fol1owing passage:

However, I think that we can quite easily distinguish between idols in whichiconism, the impulse to depict resemblancc, is thematic, and those in whichiconism is non-thematic, as in litholatry, pure and simple. So I rephrase thequestion; given that we are dealing with that dass of idols in which iconicdepiction of the object of veneration is thematic, how are we to indicate, bymeans ofpictorial 01' sculptural 'mimesis', that the artefact is endowed withanintentional psychology? There is a simple answer to tÍlis: we cannot. There isnothing physical that we can imitate here, there is no min;rhl·~bjective formthat we can copy and insert into the appropriate place in the ido!. No mattel'hGWrealistically we imitate the outward appearance of the body, we fali shortofdepicting the soul which, however, we are determined to imitate in somefashion.

To say that there is no ideal solution is not to say that there are no half-measures. Even ifwe cannot depict the mind, we can at least depict the possib-ility that there isa mind we cannot depict. By way ofa thoughte}:periment, lerus·postulate an 'ide~.lanic{)njcidol'-a sphere of perfectiy'hoo;ogéneolls mateI'''ial, actually ~I.ackbasalt. We may suppose that the spherical stone idol has amind, intentions, sensibilities, etc. but there is nothing about the material char..acteristics of the sphere, as such, which articulates with these beiiefs, which areentireiy theological and abstract. But let us modify the spherical idol somewhatby drilling a hole in it, 01' maybe two holes, which would then probably be seenas '!j'Cll'. Once the sphere was equipped with 'orifices' ofthis kind it would bepossible, not just to imagine, abstractly, that it lÍad a mind, perceptions, inten-·tions, etc. but to attach these imaginings to the formal contrast between thee,J(teriorof the sphere, into which the holes were drilled, and the interior, towhich these holes give access. Adding features which apparently make thesphere more 'anthropomorphic' (by the addition of eyes, a mouth, etc.) do not'just serve the purpose of making the sphere a more realistic 'depiction' of ahuman being, they render it more spiritual, more inward, by opening up routeso/ access to this inwardness. The 'internalist' theory of agency (in its informalguise as part of everyday thinking) motivates the dêvelopment of 'representa-tiona!', if not 'realistic' religious images, because the inner versus outer, minJversus body contrast prompts the developmentof images with 'marked' char-acteristics of i~wardness versus outwardness. Paradoxically, the developmenlof idols which depict the visible, superficial, reatures of the human body makepossible the abduction of the 'invisible' mind, awareness, and will from thevisible image. The more materially realistic lhe image, at least in certain keyrespects, the more spiritually it is seéíi.

It would be misleading to suppose, though, that the need to articulate visu-ally the contrast between inner mind and outl~rbody leads ineiuctably towardsrepresentational art forms, though this happens. My argument is that theindexical form of the mindjbody contrast, is primordiaily spatial and concell/ricthe mind is 'interna!' enclosed, surrounded, by something (the body) that is

non-mind. Now we begin to see why idols are so often hollow envelopes,with enclosures, like the hollow mau'; stones, 01' the hoilow sorcery-imageswe encountered in the preceding discussion or Polynesian idolatry (above,Sects. 7.5-6).

It is often the case that the human body (with an implied interior indicatedby orifices) is used to index this primordial inside-outside relation. But thereare other ways of achieving this as wel!. Suppose, instead of drilling 'eye'holes in the spherical idol, we leave it as it is, but place it in .'1. box, an ark.At this moment it becomes possible to think of the spherical idol in a differentway; we can easily ~uppose that the stone inside the box is the locus of agency,intention, etc. and the ark is the sacred 'vessel' which, body-like, contains andprotects this locus of agency. Once the idol is in the ark we have, once more,the physical configuration necessary for thinking of the stone as 'opposedto' something else in the way that the mind (interior) is opposed to the body(exterior). The 'homunculus-effect', in other words, can be achieved withoutãí1thropomorphizing the index, so long as ~.hecrucial ~t:ature of concentricityand 'containment' is preserved.

There are thus two basic strategies for converting (conceptually) stocks andstones into quasi~pirs~~s in artefact-form. The first of these strategies consistsof animating the idol by simply stipulating for it a role as a social other. Thesecond consists of providing it with a homunculus, 01' space for a homunculus,01' turning it into alu)munculus within some larger entity. I shall discuss the'interna!' animation ofidols in the next section. Here is an example which givesus a good view of the externalist strategy in action. Contrary to what one mightexpect, the most important images of the gods of ancient Egypt were not themonumental carved figures that have survived to this day, but much smaller,conveniently ht:ftable, idols, as the following description makes dear:

in the temple of Hathor at Denderah, there were, amongothers, the followingsacredstatues:Hathor, painted wood,copper, inlaideycs,height 3 ells,4 spans, and 2 fingers;Isis, painted acaciawood, eyes inlaid, height 1 eU;Horus, painted wood, inlaid eyes,height 1 ell and [ finger.The largest, therefore, was scarcelyof life size; the smallcstonlyabout 16 illches in height. The reason for this insignificancein size was that for :,'certain acts of worship the imageshad to be easilyportable.

The paltry size and material of these little woodendolls were, however,atoned forby the splendour oftheir abade, and the reverencewith which they were served.TheshrÍllê 'of the god was in the innermost chamber of the tempie, which was in totaldarkness saveon the entry of the officiatingpriest bearing artificiallight. It consistedgenerallyof a sillgleblockof stane, often, especiallyin the later periods, of enormoussize, hewn into a house which surrounded with impenetrable walls the imageof the

gorl. The doorway in front was closed with bronze doors, or doors of wood overlaidwith bronze or gold-silver alloy ... after the daily ritual had been gone through, thesedoors were closed, fastened with a bolt, and then licd with a cord bearing a clay seal ...Within the shrine, the image of the god reposed in a little ark, or portable inner shrinewhich could be lifted out and pl~ced upon the barquein which the deity made his jour~neys abroad on stated occasions.

The daily ritual of servicc to the image was in ils main outlines the samc in ali lhetemples .. '. the procedure was as follows. Early in the morning the priesl of the day,after lustratlOns, entered lhe Holy of Holies, bearing incense in a censer and stoodbefore lhe shrine. He first loosened the door thal closed the shrine, repeati~g as he didso a stereotyped phrase: 'The cord is broken, and lhe sealloosened,-I come, I bringthee the eye of Horus [i.e. light, the sun]' ... As lhe doors of the shrine opened andthe god was revealed, the priest prostrated himself and chanted 'The gales of heavenare opened, and the nine gods appear radianl, lhe god N is exalted upon his grealthrone ... Thy beauty belongs to thee, O god N; thou naked one, clothe thyself.'Taking his vessels, the priesl then began to perfi)Im lhe daily toilet of lhe god. Hesprin~led water on the image lwice from four jugs, clothed it with linen wrappingsof whlte, green, red, and brown, and painted it with green and black paint. Finally, hefed the image, by laying before it bread, heef, geese, wine, and water, and decoratedits table with tlowers. (Blaikie 1914: 132)

It is not hard to see the applicability of the externalist theory to the cult ofthe idols in the tempIes of ancient Egypt, cited above. The daily round towhich the idols were subjected, being woken in the morning, washed, made up,served breakfast, and so on, imposed agency on them willy-nilly by making thempatients in social exchanges which imply and confer agency necessarily. Thereis no 'as if' or make-believe about such performances; they would be point!essunless these life-endowing rituaIs were literal transpositions of the means inwhich we induce agency in social others in human form, such as children.

Indeed, it is very hard to read this description without being reminded ofchildren's play with dolls. This is not altogether an appropriate comparison,except to the extent that children do not 'play' with dolls but actually makea cult of, or worship them. 'Play' behaviour is supposed to take place in cOll-ceptual brackets, which say 'this is pIay-so I am not doing what I appear tobe doing' (Bateson 1936). Children, outside the temple,might play at beingpriests, and pretend to worship toy gods, but this type of make-believe isentirely distinct from the activities of the priests themselves. They were notat play, but at work. They were serious.

~or is it quite right to say that their actions were '~ymbolic'-though ofcourse everything rather depends on how the word 'symbolic' is understood.Offering food to the image orthe god is not a pantomiirieor dumb-sliôw, as ifthere were some alterna tive way of feeding a god which was being alluded to,but not performed. Receiving food offerings is how the Egyptian gods ate theirfood. This is not to say that the act of feeding the god by placing an offer-ing before it is not symbolic in the sense of 'meaningful', but the 'meaning'

stemmed from the real (causal) outcome of this act of feeding; the god was nolonger hungry. The essence of idolatry is that it permits real physical inler-act;ons to take placebetween persons and divinities. To treat such interactionsas 'syinbolic' is to miss the painl. Images can be employed in worship in non-idolatrous ways, as aids to piety but not physical channels of access to thedhririity-the Christian use of religious images is supposed to belong to thiscategory, ihough in practice many Christians' use of images is de jàclo idola-lrous, if not admitted to be so. We can only distinguish between idolatrousand non-idolatrolls use of religious images because idolatry is in an importantsense not 'symbolic' at alI, whereas the use of images as aids to piety, ratherthan physical vehicles of divinity, is syrnbolic. The Egyptian ceremonies justadduced belong firmly in the category of idolatrous practices, and are thus real,practical, services performed for divine social others in image-form, not sym-bolic acts.

Ali the same, the way in which these idols were enmeshed in the structuredroutines of daily life can only provide a partial answer to the problem of idol-atry. There are other features of the situation which seem to indicate the oper-ation 'õf different factors, which cannot be accounted for in terms of a purely'externalist' notion of agency. For instance, as noted above, the pureIy extern-alist tlteory of agency makes no stipulations as to the physical or visual formof the object (index) which is treated as an agent. These Egyptian idols were,in fact, quite realistic representations of the outward appearance of humanpersons. The externaIist theory of agency is not in a position to difTerentiatebetween 'iconic' idols, such as these, and 'aniconic' idols, such as stones, orthe planks of wood, known as haitulia, which the Greeks worshipped (asAphrodite, Zells, etc.) before, and alongside, their subsequent cult of sculp-tural images of these gods. This indifTerence towards the iconic properties ofidols is, in a sense, a point in favour of the 'externalist' theory; unless agencywere a purely externally endowed property of idols, unconnected with theirphysical substance or form, then it is hard to see how the worship of aniconicidols such as stones or planks of wood would be possible. The externalisttheory has got to be at least half-right, for this reason. Yet it cannot be whollycorrect, or the impetus towards 'shaping' the idol-not just treating it as anagent, but making it look like or share physical attributes with a 'prototype'-would be inexplicable. Here one has to introduce the other theory of agency,the 'internalist' or homunculus theory. While it may be true that the agency ofthese idolsderived in partfrom the way in which they were inserted into therelational texture of 'external' social praxis and language, this 'passive' agencyis certainly not the whole story. These particular idols were in fact highlyiconic, and moreover, the description given (by Blaikie, but deriving directlyfrom Herodotus) ernphasizes particularly t~ir 'inlaid eyes'. In the next sectionI will describe in detail the consecration of contemporary idols, which cruciallyinvolves the animation of images by providing them with eyes. Eyes are, of ali

body orifices, those which signify 'interiority' (i.e. the possession ofmind andintentionality) most immediately (see the prcvious discussion of darshan). Theparticular attention paid to the eyes of thesc idols arises, not from the need torepresent the body realistically, but from the need to represent the body insuch a way as to imply that the body is onú' a body, and that a much moreimportant entity, the mind, is immured within it.

Thus, it is equalIy important to note that the animation of these idols wasbei:lg achieved, simultaneously, in a quite different way. The description givcnabove emphasizes, besides the realistic form of the idols, their extraordinaryand impressive surroundings. They were kept, except when being served bythe priests, in a box or ark, which, in turn, was kept in the darkest and mostcentral sanctuary of a vast temple complex, consisting of innumcrable lessersanctuaries, shrines, courtyards, barracks and workshops, etc. If we situateourselves, not inside the innermost sanctuary, but outside in the courtyard,with the ordinary worshippers (who rarely if ever saw the idols themselves)then we may readily imagine that the idols (immured in the temple complex,and animating it like a giant body) come to stand for 'mind' and interiorit~not just by physical resemblance to the human body, but by becoming thtanimating 'minds' of the huge, busy, and awe-inspiring temple complexo Justâs the 'mind' is conceived of as an interior pcrson, a homunculus, within thebody, so the idols are homunculi within the 'body' of the templc. And it istrue that idols, even very representational idols, are invariably presented in asetting, a tempie, a shrine or an ark, a sacred space of some kind, which hasthe effect of emphasizing their interiority, their secludedness and (relative)inaccessibility, as weIl as their majesty. The seclusion of the ido] has, auto-maticaIly, the effect of motivating the abduction of agency, on the basis of theequation:

idol : temple :: rründ : body.

These reflections lead towards a relativization of the contrast between theexternal and internal conceptions of agency, sentience, etc. with which I beganthis section. Ir is obvious that the homunculus, or 'inner person' conception ofagency essentialIy reduplicates, within the human person, the relation whichalways exists between a human person and a texture of external relationships,but within the interior domain, within the hody. This imagery leads to the'homunculus within a homunculus' problem which besets this type of theory,according to its critics. But this'problem' is also an advantage, in that it tendsto blur the distinction between the 'induced' kind of animacy which is imposedexternalIy on the idol by enmeshing it in praxis, language, social relations androutines, and the 'interna!' agency which the idol is supposed to possess as a'mind' encapsulated in a surrounding body. Just as the idol, externally, is at thecentre of a concentric array of relations betwecn persons, so the idol, internalIy,can be seen as a concentric array of relations between the 'inner' persons-

the pandemonium of homunculi--of which it is composed. Let us turn to anexample which reveals this in a particularly graphic way.

7. I I. Concentric Idols and Fractal PersonhoodTo exhibit the animation of the idol through the congruence between theexternal relational context within which the idol is set, and the internal nexusof relations between the mind and the body (as a relation between inner andouter 'persons') consider the Polynesian example in Fig. 7.I1{[, a carving fromRurutu in the Austral Isles, Which has been in London since [822 and which'can be seen at the Museum of Mankind. This carving, in Rurutu called A'a,but more commonly identified as Tangaroa, is arguably the finest extant pieceofPolynesian sculpture. Almost every other Rurutan idol was consigned to theflames by the missionaries, but this one was preserved, initially to drum upsubscriptions for the London Missionary Society so that they could afford todestroy other, no doubt equally fine, carvings elsewhere.

The most striking attribute of this carving is the way in which the featuresof the god are represented by little figures which repeat, in miniature, theoveralI form of the god as a whole. This god sprouts Iittle gods alI over itssurface: mathematically, it is akin to the type of figure known as a 'fractal', afigure which demonstrates the property of self-similiarit)' at different scales ofmagnification/minification. Moreover, besidesbeing a god made ofmany gods,the A'a is also a box or an ark. Ir is hollow inside, having a lid at the back,átid ir originally contained twenty-foúr or more additional, smalIer images ofRurutan gods, wtiích were removed and destroyed in 1822. For ali we know,the gods inside lhe A'a were themselves holIow, though I think noto Butwhereas we think ofboxes as less significant than their contents, the A'a, eventhough it is a box, is the primary image ofRurutan divinity, encompassing andsubordinating alI the subordinate gods who sprout from its surface and onceresided in its interior. According to contemporary Rurutan traditions, theexterior gods encompassed by the A'a correspond to the kinship units (e1ans)comprising Rurutan society as a whole.-.j-~any other important Polynesian .~.

, Curiously though, contemporary Rurutans, according to the ethnographer Alain Babadzan have "quite ditTerent theory abOut the gods which, they know, were once inside the Na. According ~o th~Rurutan elders, there were three gods inside the A'a when it was made, by a Hero named AmalteraLAmaiterai made ihe A'a after visiting a city none other than London, present resting-place of the A'a,which he reached in fulfiiment of a species of knightly quest, imposed on him in order to win the handof the adopted daughter of the King of Rurutu, who had been promised to his b~other. In LondonAmaiterai encountered the God of Wisdom (who later was the God of the Chmtlans, brought toRurutu by the missionaries) whose image he replicated in the form of the fam~us A'a. The gods insidcthe A'a were three Polynesian gods originating in London: Room-etua-ore, ahas Te Atua Metua, ahasGod the Father; Aura-roiteata, alias Te Atua tamaiti, alias God the Son; and Te atua aiteroa, alias TeAtua Varua Maita'i alias God the Holy Spirit. In other words, the A'a is the Tabernacle in which theTrinity arrived on Rurutu, by the agency of a Rurutan hera, long before the missionaries themselvesarrived. The A'a is in London, but it is present on Rurutu in the form ofChristian belief.

FIG.7.I1/J. The fractal god: A'afrom Ruruta. Source: The BritishMuseum, MM 011977

carvings represent personhood in the form of genealogy, as for instance therei"ated carving, a 'staffgod', from the Cook Islands shown in Fig. 7.11/2.

What is particularly remarkable about the A'a is the explicit way in whichthis image of a 'singular' divinity represents divinity as an assemblage of rela-tions between (literally) homunculi. In so doing, the A'a obviates the con-irast between one and many, and also between inner and outer. The surfaceof this image consists of amalgamated replications of itself, or alternatively, asuccession of budding protuberances. Internally, the image consists of itself,replicated on a smaller scale, within its own interior cavity. As such, it images 'both the notion of personhood as the aggregate of external relations (the out-come of genealogy, fanning out in time and space) and at the same time thenotion of personhood as the possession of an interior person, a homunculus,õr; in this instance, an assemblage of homunculi. We cannot individuate theA'ain the way inwhich we normally individuate persons by identifyingthe boundaries of their person with the spatial boundaries of their bodies, forthe A'a has no such boundaries; it is Iike a Russian doll, and in this respect,it irresistibly recalls the lines in Peer 'Gynt in which the hero compares the(moral, biographical) person to an onlon, composed of a succession of concen-tric layers:

FIG. 7. II/z.Genealogicalpersonhood objectified:a staff god from theCook Islands. Sourcr:Museum of Archaeologyand Anthropology,Cambridge, Z 6099

---Why, you're simply an onion-and now, my good Peter, I'm going to peel youand tears and entreaties won't help in the least.ITaking an onion, he strips it skin by skin)There goes the battered outer layer-Ihat's the shipwrecked man on the dinghy's kccl.This layer's the passenger-scrawny and thin,But still with a bit of ataste of Peer Gynt.Next underneath comes the gold-mining Self-the juice, if it ever had any is gone.This rough skin here, with the hardened patchis the fur-trapping hunter from Hudson's Bay.We'I! throw that away without a word.Next the archaeologist, short but vigorous;and here's the prophet, juicy and fresh-it stinks of Iies, as the saying go~s,and would bring tears to an honest man's eyes.This skin, curled and effeminate,is the gentleman living his life of pleasure.The next looks unhealthy and streaked with black-black could mean either priests or niggers ...lHe pee/s ojJ severa/ /ayers at once.)What an incredible number of layers!Don't we get to the heart of it soon?

[He pu//s the whole onion to piem.]No, I'm damned if we do. Right down to the centrethere's nothing but layers-smaller and smaller ...Nature is witty!

from an inner subjective self, is to be sought in this 'enchainment', the struc-rural c(}ngruence between thetnner self (which is relational) and tite outersefí (which is equal1y rdarional, but on an expanded scale). The 'genealogicaltheory of mind' which is explored, particularly, in Strathern's work (1988)seems perfectly expressed in the forro of the A'a.

But this artwork, and its Polynesian cousins, is not unique. In fact, thepos~~~,i~rt0f 'significam interiors' is a verycommon feature of sculpt~ral~~rl{sspecifically intended for cult use, rather than as mere representations.Sculptural images which open up, like the A'a, to reveal other images, weremanufactured in ancient Greece, and provided Alcibiades with a simile for hismentor, Socrates (as narrated by Plato in the Symposium 213-15):

I am here to speak in praise of §2.c.rates,Gentlemen, and I will just do it by meansofsimiles. Oh yes, he will perhaps think it is only for a bit of fun, but my simile willbe for truth, not for fun. I say then, that h~,is exactly likea Silenos, the little figurest,haryouseein statuaries' shops; the craftsmenmakethem, they hold panpipes01'pipes,and t!t~ycanbe ()penec!.IlPdown the middle01'foldedback,and then they showinsidethem, imagesofthe gods.And I say further, that he is likeMarsyas the Satyr [whowasflayedby Apollo] ...

Here the contrast is between the ugly exterior (body) and the divine interior(mind) of Socrates. The same kind of image was also deveIoped in Christianêiílfárt, though with a different theological implication. A c1assof holy statuescalled 'v!~r.Kt:!!.9\lvrantes'was made in the Middle Ages, though fewhave sur-vived into the present, perhaps because these images were particularly con-ducive to idolatry, as the following passage from Camille's work of The GothicIdol shows. 'Our Lady ofBolton' which used, before the Reformation, to standin a chapel of Durham cathedral was:

a marveylousIyvelyand bewtifull Image of the picture of our Ladie, so calld Lady ofBoultone, which picture was made to open with gymmers [01' two leaves]from heI'breasts downwards. And ~ithin this saide immage was wroughte and pictured theimmag:eof our Saviour,merveylousefynliegilted holdyngeuppe his hands, and hold-íng betwixt his hallds a fair largeCrucifixof Christ, ali of gold, the which Crucifixwasto be taken fourthe every Good Fridaie, and every man d'idcreepe unto it that was inthat Church at thar daye. And every principall daie the said image was opened, thatevery man might sce pictured within heI' the Father, Son and the Holy Ghost, mostcuriouslyeand fynclygilted. (Camille1989:230-1)

Camille iIIustrates a German example of such a 'vierge ouvrante' (Fig. 7.11/3).'Our Lady of Bolton' was related to the more common type of Christian holyimage in the form of a r~.Iiquary,in that she was, besides an image of the Virgin,a coIltainer for the golden crucifix which was paraded on Good Friday. (Foran l!1dian parallel to a 'vierge ouvrante' cf. Fig. 7.11/4 showing Hanuman,the monkey god, opening his breast to reveal Rama and Sita.) Such images"lere controversial even when they were still in common use: Camille quotes

Peer Gynt's onion is also a fractal, of the same essentially concentric form asthe A'a. Ibsen's idea, in utilizing this image is to show that there is no ultimatebasis to Peer Gynt's personhood; he is made of layers of biographical (reh-tTonal)experience accreted together, for which none the less, he must take soleresponsibility. Perhaps it is not such a vast step to pass from Peer Gynt trappedin the aporias of nineteenth-century materialism and individualism to the theo-logical impulse which motivates the A'a, whieh depicts the divine ereator, rhemind of which the world is the body, in the form of a body composcd of otherbodies, ad infinitttm.

This idea is given contemporary expressionin the work on personhood inMelanesia, by writers such as Marilyn Strathern (1988; cf. Ge1lÍ998) and RoyWagnér(I991). Wagner, in particular, has developed the notion of 'FracralPersonhood', which he mobilizes to overcome the typical1y 'Western'opposi-tions bet'\Veenindividual (ego) and society, parts and wholes, singular and plural.The notion of genealogy, which is so signally expressed in our two Polynesian'examples (both idols, of course), is the key trope for making plurality singularand singularity plural. Any individual person is 'multiple' in the sense ofbeingthe precipitate of a m'i:lititude of genealogical relationships, each of which isi·nsta.ntiatedin his/her person; and conversely, an aggregate of persons, such asa lineage 01'tribe, is 'one person' in consequence of being one genealogy: theoriginal ancestor is now instantiated, not as one body but as the many bodiesinto which his one body has transformed itself. Wagner writes:

A fraetal person is never a unit standing in relation to an aggregate,01' an aggregatcstanding in relation to a unit, but alwaysan entity with relationshipintegrallyimplied.Pe1hapsthe most eonerete iIIustrationof integral relationshipcomes from the gener-alised notion of reproduetion and genealogy.Pcople exist reproduetively by being:'earried' as part of another, and 'earry' 01' engcnder others by making themselvcsgenealogieal01' reproductive 'factors' of these others. A genealogyis thus an enehain·ment of people, as indeed persons would be se(:n to 'bud' out of one another in aspeeded-up cinematiedepietion of human life. Pcrson as human being and person ai,lineage01' c1anare equallyarbitrary seetioningsor identifieationsof this enehainment..different projeetionsof its fraetality.But then enchainment through bodily reproduc-tion is itselfmerelyone of a number of instantiationsof integral relationship,which isalsomanifest, for instanee, in the commonalityDf shared language. (1991: 163)

From the anthropological point of view, if not the philosophical one, th(:solution to the conflict between the external notion of agency, deriving frominsertion in the social milieu, and the 'internalist' theory of agency, deriving

FIG.7.1I/3. 'Viergeouvrante'. Source: C.~mílle1989: pl. 124. NurembergGermanísches NationalMuseum. Painted woodH. 126 em

FIG. 7.11/4. An Indian paralldto 'vierge ouvrante': themonkey god Hanuman revealsRama and Si ta in his breast.Sourte: A. Mookerjce (1980),Ritlla/.4rt ofJudia (London:Thames & Hudson), plate 80.Kalighat sehool, southCalcutta, LI 850. Gouacheon paper

the theologian Gerson, in 1402, denouncing 'Carmelite' images which 'hayethe Trinity in their abdomen as ifthe entire Trinity assumed flesh in the VirginMary ... ' (1989: 232). These images were ali too animate, they smacked ofnecromancy rather than religion. Later on in the same book, CamilIe sho\Vshow reliquary heads, containingthe bones af saints, could become objects ofdeep ~fficial suspicion, especialIy the silver and golden heads supposedly wor-shipped in pagan ceremonies by the Knighls Templars, who were supprcssI:din 1308as a result (ibid. 271--']).The notorious 'talking head' reputedly deviscdby Roger Bacon is another variation on the same i:heme (ibid. 246---'7). But lherelation between religious art and sorcery runs very deep, we have already secn.

No such opprobrium attached to more ordinary reliquaries in human form, ofwhich there are many famous examples, sueh as those illustrated and discusscdby Freedbe~g (1989: 92-5, figs. 30-2). This author also remarks on the filetthat in the Middle Ages, churches could not be consecrated at alI unless theyhad holy relicsinstalled in them. Just as rdics animated the re!iquary image,rendering it a particularly holy object (an object with a mind, or perhaps moreprecise!y, a spirit, within it) so the church (fabric) as a whole became a 'body'whose animation also required the insertion of a relic. The normal place for lhe

insertion of relics into the fabric of a church was inside the altarpiece, in thealtar itself, or buried beneath it.

The insertion of animating relics into images raises a fresh question, how-ever, that of consecration-the management of the transition between thereligious image as a 'mere' manufactured thing and a vehicle of power, capableof acting intentionally and responding to the intentions of devotees. We shallconsider this aspect of idolatry in the next section.

7.12. The Rites ofConsecrationIr is fair to say that the worship of images or idols is most extensively prac-tised, nowadays, in south Asia, among Hindus and (in a rather more qualified

way) among Buddhists. Here images are still being produced and installed inholy places in great numbers and here it is possible for anthropologists toobserve the rites of consecration in dctail. I shall therefore conduet the argu-ment of this section with reference to three well-known south Asian examples,the consecration of the idol of Jagganâthin Puri (tschmann tt ai. 1978), theaccount given by Richard Gombi-ich (1966) ofthe installation of a statue oftheBuddha in a Sinhalese monastery, and finally, the aecount given by MichaclÃftên (1976) of the: consecration of the 'Iiving image' of the goddess Talejll(purga) in Kathma.ndu.

The images of Jagganath, 'his brothers, Balabadhra and Subadhra, and hiswife Sudarsana in the tempie ofPuri (Orissa) are among the most revercdidols in the whole ó( IIlala~but, though anthropomorphic, these images are(visually speaking) quite obviously cylindrieal sections of tree-trunk, dresscdup and equipped with vestigial upper limbs and very, very, large eyes. Theimages are not old in physical terms, though their design is indeed anciem,sinee they hãve to be renewed every twelve, or at mostevery nineteenyears,in the eourse of a eeremony called navakalevara. A detailed description ofthisceremony (eont~ibui:ed by G. C. Tripathi) is provided in Esehmann'swork on the history and affinities in 'tribal' re1igion of the Pu ri Jagganath cult.The tribal affinities of this importam Hindu eult are not in doubt; I mysclfworked in the same region ofIndia (Bastar distriet, which abuts onto the Orissahighlands to the west) among tribal Muria Gonds, whose images of divin-ities were aniconic wooden posts, without limbs or eyes, but otherwise highlyresembling the Jagganath images, though on a reduced scale (A. Gell 1978; d.S. Gell 1992).

Tripathi's account of navakalevara is a first-rate exercise in ethnographicdescription, whieh I cannot unfortunately sllmmarize in any detail. The cerc-mony of renewal has five phases, as follows:

r. To find out the daru (sacred wood) with the prescribed eharacteristicsand to bring it to the temple (involving a sacrifiee to ward of evil spiritsete. and to sanctify the tree before felling it);

2. The carving of the wooden strueture of the images;3. The conseeration of the images by the insertion of the 'Iife-substancc'

(brahmapadartha) inro them;4. The burial of the old figures, the funeral and purificatory rites of the

Daitas (temple servitors, of low caste);5. Giving the images their final form by rneans of several coverings of c10th

etc. and by applying paint to them (Eschmann 1978: 230).

I omit the first two phases in this process, cxcept to mention that the efficacyofthe images depends, initially, on the auspieious location and form ofthe darutrce (it must grow by water, be surrounded by thrce mountains, have dark,-'red' bark, a straight trunk, with four branehes, ete.). Numerous ceremonies

accompany the felling, transport, and carpentering of the daru tree. The imagesare made by tempie servitors of low caste but high ritual privilege, the Daitas,ex:.tí-ibalswho alone know how to create this type of wooden image, accordingrõ precise rules, importam aspects of their work being kept secret by them,even from the Brahmin priests. Among the most important of their secrets isthe actual nature of the 'life-substance' (brahmapadartha) of the images.

The consecration procedure commences simultaneously with the carving ofthe images, and is first of ali conducted by Brahmins. Since the images are stillbeing made, the Brahmins devote their efforts to consecrating a separate pieceof daru wood, which, divided into four pieces, will become the 'Iids' over theêãvities in the finished images containing their 'Iife-substance'. This piece of~()()d -is protected with offerings to evil spirits, then elaborately bathed andpurified, then placed to rest on a ritual bed. After this, the spirit of the godNarasimha (of which Jagganath is a form, as well as being a form of Krishna)is induced into the wood by the recitation of mantras on each of its parts(eq~ated with the parts of a body) over a number of days. This invocationprocedure is called nyasadaru. After this, the ritually treated daru wood-andby metonymy, the larger sections of daru wood which are at this momentreaching complction as images in the Daitas workshop-is spiritually speakingcndowed with !ife, flesh, blood, sense organs, etc. It can thcn be eut imo fourlid-pieces to fit into the four images.

The crucial ceremony however, is not eondueted by the Brahmins but bythe Daitas. Thcse take the 'old' images from the temple, and strip from themtheinany layers of resin-impregnated cloth with which they are bound. Theycãfi then reach the eompartments inside the old images in whieh their 'life-substance' is secreted.The Daita entrusted with the job opens the belly of the old image in dead of night withhis eyes blindfolded and the hands wrapped up to the elbows so that he mayneither see nor feel the brahmapadãrtha of the image. The casket containing theBrahmapadãrtha is then taken out of the old murti [image] and placed m the new one.The cavity ofthe new image is then covered with one ofthe four pieces ofthe Nyasdaru[sacred wood] ""hich has been consecrated for about two weeks by the Brahmins.(Tripathi in Eschmann 1978: 260)

Nobody really knows, except the Daitas, exactly of what object or substancethe brahmapadartha is composed. It may be a relic, a portion of a woodenJagganath supposedly incinerated (but not completely) by the Muslim icono-clasts under Kala Pahada in 1568. In the opinion of the Brahmins, it is asalagrama, a type of sacred stone, usually an oval river-pebble from the Him-alayas containing fossilized ammonites, and/or cavities (see Fig. 7.12/1 fromMookerjee and Khanna 1977).

The Daitas now perform burial rites over the deeeased images which havelost their life-substance. They weep and mourn, observing mortuary pollutionrestrictions for ten days. However, they derive benefit from this, in that they

FIG.7.11./1. 'World-seed':salagrama. Source: Mookerjecand Khanna 1977

this purpose. After this, the new images are paraded and installed in their placein the temple with great pomp and ceremony (Tripathi, ibid. 262-4).

Two aspects of these intel'esting ceremonies are of particular concern to usoThe consecration of the images evidently proceeds according to two parallel~!r~~egies simultaneously; first of ali, the strategy of the Daitas, which focusesparticularly on the placing of the Iife-substance inside the image, in the cavity,~nd secondly, the strategy of the Brahmins, which proceeds in the inversedirection, through tite apostrophization of the representative billet of wood with~ife-endowing mantras (the phase of nyasadaru) and the application of the final~troke of paint to the pupils of the images' eyes. In other words, the Daitas' pro-cedure could be called the inside-out procedure, while the Brahmins' procedureis the o~ide ...iJ1 procedure. Both are necessary and mutually complementary.

~E.~~aitas' procedure, is, perhaps, the more primitive, in that it seeks torender a physical analogy between t~e possession of a 'soul' and the possessionof an interior cavity inhabited by a homunculus. Moreover, it is 'genealogical'because it establishes a kinship link between the old images and the newones through the transmission of substance between 'generations' of images.~~e. Brahmin strategy, for the most part, is more abstract than this; ratherthan fashion the image, they address it with ma.ntras, animating it externallythrough 'the magical power of words' (Tambiah 1985). We can easily see thatthese two strategies correspond exactly to the two strands previously identified

. in the philosophy of mind and agency. The Daitas' strategy is the 'internalist'.?~:'..the Brahmins' the 'externalist' oné. But it would be false to supposethat these two strategies are independent. In the end, the Brahmins concede tothe internalist strategy to the extent that their final act is one of physical modi-ncatíon of tht: image, not mere apostrophization. Their culminating act ofp.ainting in the pupils of the eyes of the image is mimetic and iconic. Thoughthe eyes of these images are not actually transparent, the pupils of any eyes arenever 'things' but always ho/es, orifices, giving access to the hidden interiorwithin which 'mind' resides. The surface of an idol is not an impermeable bar-rier, but a means of access to this essential interior.

In fact, the images of Jagganath and his companions are a series of 'skins'just like Peer Gynt's onion, with the same implication that the ultimate centrecan never be reached. Tile outermost (relevant) skin of the idols consists of thetemple of Pu ri itself (which is, of course, a microcosmos) which is filled withsacred words and odours-verbal and olfactory skins (cf. Anzieu 1989: 59).The idols reside in the centre of, and animate, this reverberating microcosmos,and are animated by the incessant flow of sacred words. Proceeding towardsthe centre we approach the idols through a 'social skin', the throng of pilgrimsand attendant temple servants and priests, who, by their attentions and devo-tions, animate the idols occupying the cynosure of a great assembly of souls.The idols themselves are enshrined at the centre, framed on their altars,adorned with masses of flowers and jewellery, presiding over their material

are allowed to keep the cloth wrappings of lhe old images, which they cut intostrips and sell to pilgrims. These pieces of cloth confer protection and güodfortune on the purchasers.5

The final phase of consecration now oecurs. Another caste of tem pie ser-~ants, said to be Kayasthas (a low form of Brahmins) undertake the wrappingof the images-who, in wooden form are just 'skeletons'-with their 'flesh'th~t is~ with cloth an~ paint. The wooden 'bones' are first of ali '\y.ashed' b;bemg Impregnated wlth camphor oil. This gives the bones 'marrow'. Then,apparently, long red threads are wound around them, representing 'bloodvessels'. After this, many strips ofred cloth (for flesh) impregnated with resin(blood) and starch (fat, semen) are added, till the image begins to assume itsfinal formo The ou ter wrappings are the image's skin.

Finally, the images are p_ainted by craftsmen skilled in this art (chitakara).The very last act, which finalizes the consecration of the new images, is lhepainting in of the pupils of the immense eyes of the images, which is done bythe Brahmin priests thtmselves, reciting Vedic mantras the while. After gi vingthe last stroke of paint to the eyes of the images, the Brahmins give each a balh,to remove the pollution from the previous contact the images have had withlow-caste carpenters, painters, etc. This is done not by bathing the imagedirectly, but bathing the images' images, cast in large bronze mirrors kcpt fi>r

, This prerogative recalls the 'feather exchange' following the decortieation of the Tahitian to 'o I'i.6abo:e). The mana of Jagganath is disseminated via his exuviae, his body-parts, jutha (Ieavings) ele.Incldentally, the same is true in the !'.gyptian example discussed in 7.10 above. Every time the idols'c10thes were renewed. the discarded garments were distributed to important pcrsons, to use to li rapthe.r own corpses m as mummy-c1oths. Whereas in PQlynesiaand India, the c1oth/feather exuviae of lhegod wenl to benefit the lil'ing, in Egypt they benefited lhe dead in the afterlife.

wealth of heaped-up offerings, their external skins in the form of possession.\.V/e cannot approach 01' touch the idols, so we can proceed further on our ;OUI'-

ncy towards the centre in imagination only. What we see are their visible skins,but these are only outer wrappings. These wrappings eonsist of numerollslayers, their inner skins of flesh, fat, semen, blood, bone, marrow. These wemay penetrate, one by one, 01' gazing into the enormous eyes of th~ idols, wecan enter their bodies directly. But what is there, concealed beneath, behind,the inky pools? Inside, there is a primordial cavity, an internal skin. And insidethe cavity, an animating presence of some kind. Conceptually, we know that,in fact, this cavity contains a casket, another inner-inner skin. And what is inthe casket? Even the man who placed it there does not know perhaps; he hascertainly never seen its contents. We believe that there is asalagrama there, a~acred stone 01' world-seed. If so, what is in the salagrama?'The salagrama itselfhas an interior, and, holes leading into this interior (cf. Fig. 7.12/1). We mustenter these holes. And then what would we find? Who ean say-and does il:matter?-for by now it is apparent that the animation ofthe image is not a mar.•ter 01' finding the 'saered eentre' at ali. What matters is only the reduplicatiollof skins, outwards towards the maeroeosm and inwards towards the micm.cosm, and the fact that ali these skins are !;trueturally homologous; there isno definitive 'surface', there is no definitive 'inside', but only a eeaseless pas-~age in and out, and that it is here, inthis traHic to and fro, that the mystery 01'animation is solved.

The eonseeration of an image of the iBuddha in Sri Lanka, deseribed bvGombrieh (1966; ef. the diseussion in Freedberg 1989: 84-7,95) follows a sim:.ilar pattern, though redueed. Buddhists, especially monks, are not supposedto worship idols, but showing respeet to images of the Buddha, by making~fferings and gestures of submission, is one way to aequire merit and secure agood rebirth, if not a very significant one. Otlee again, it proves to be the casethat images ofthe Buddha are, ifthey are to have any religious importanee, alsoreliquaries. Minute portions of the Buddha's bodily remains are plaeed insideimages to render them efficaeious. However, this does not eonseerate them,aecording to Gombrieh; eonseeration is accomplished by the eraftsman, whopaints in the eyes of the image in the course of a speeial eeremony called netrapinkama the 'eye eeremony' (1966: 25).

The eeremony presents some interesting comrasts to the eonseeration of theJagganath images, as well as many points 01' continuity. The Brahmin/Daitarelationship is partially inverted in the ritual division oflabour in the Sinhale~(:Buddhist eonseeration eeremonies. In Puri, it is the Daita who animates th<:image by putting 'Iife-substanee' into it, and the Brahmin priest who paints inthe eyes; in Sri Lanka, on the other hand, it i!,the monk (abbot) who places therelie imo the Buddha statue, and the lay eraftsman who paints in its eyes. This

. is a precise refleetion of the differenee between the Buddha (a dead humanbeing with moraUy supernatural eharaeteristics) and the Hindu gods, who are

. nan-human immortals. A Buddha statue celebrates the possibility of 'a gooddeath' and monks are semi-dead individuaIs who aspire to the ultimate good-death eondition. Consequently it is only appropriate that the handling ofBuddha relies, whieh are pieees of a dead body, should be assigned to monks,who are semi-dead themselves, and who of eourse always preside over fun-éraIs (but not births and marriages) in Buddhist countries like Sri Lanka (01'

Thailand; ef. Tambiah 1985). In a sense, then, what the relie does is make theBuddha statue like the Buddha, by making it 'dead' through the insertion of a'death-substance'-in the rather paradoxical sense that Buddha-hood impliesdeath-in-life. However, one ean see that this ritual procedure hardly amountsto the 'animation' of the image. Gombrieh understands this as the means of'Iegitimating' the statue as a Buddha image, so that from the monkish per-spective no taint of idolatry arises. For the laity things are otherwise. Buddhais reaUy a god, and approaehing the statue of the Buddha, the lay worshipper,immersed in life and sin, seeks personal reassurance through eommunion withthe Buddha's living presenee, eonveyed through the eyes (see above on theHindu equivalent of this form of worship, darshan). Because only lay peopleare 'superstitious' enough to engage in sueh theological1y decried aets of wor-ship, it is the lay craftsman who is eharged with the task 01' animating theBuddha image.

The ceremonyis regardedby its performersas verydangerousand is surrounded withtabus. It is performed by the craftsman who made the statue, after several hours ofceremonies to ensure that no evil will come to him. This evil, which is the object ofall Sinhalese hcaling rituaIs, is impreciselyconceptualised,but results from makingmistakes in ritual, violating tabus, 01' otherwise arousing the malevolentattention ofa supernatural being, who usually conveys the evil by a gaze (Mima). The craftsmanp<iiritsiri the eyes at an auspiciousmoment and is left alone in the dosed tempIe withonly his colleaglles,whileeveryoneelse stands c1earevenofthe outer door. Moreover,the craftsmandoes not dare to look the statue in the face,but keepshis back to it andp~intssideways01' overhis shoulderwhilelookinginto a mirror, whichcarchesthe gazeof the image he is bringing to life. As soon as the painting is done the craftsmanhim-self has a dangerousgaze.He is led out blindfoldedand the coveringis only removedfrom his eyes when they will first fali upím something which he rhen symbolicallydestroyswith a swordstroke. (Gombrich 1966: 24-5) [This couldbe an animal,sllchas a buli, but a pot, 01' a tree whichexudes sap can be substituted.]

The detail of theinirror recalls the use of bronze mirrors by the Brahminpriests of Puri to 'bathe' the images by splashing water over their reflections.But the Puri Brahmins are unafraid to look at the images direetly while paint-ing in their eyes. Unlike the eyes 01' Jagganath, the eyes 01' the Buddha, whenfirst opened, inflict death on the very person who performs the opening (a caseofArtist-A _.~ Artist-P; see above, Seet. 3.I1). The eraftsman must there-upon slay some creature in order not to die himself. This is very far fromtheologieally Buddhist, and in olden times, the eraftsman was dressed as a

'king' while performing the ceremony, that is to say, as the negation of a monk,a violent, worldly figure, who by sacrificing religious merit himself, allowsothers (monks, devotees) to achieve it.

Freedberg (1989: 95), commenting on Richard Gombrich in his magnificenttreatise on the reception af religious and other types of images, raiscs a verybasic questiono Are images such as these powerful and efficacious (religiously)because of the ceremonies of consecration which have endowed them withsignificant, occult, characteristics, or is it because they are, first and foremost,images, linked by the power of mimesis to the deities they represent? Citing thesupport of Gadamer, he opts to assign primacy to representation. He arguesthat images work because they have intrinsic signifying-functions, which canbe separated from the kind of efficacy possessed by religious objects (such asrelics) which have not becn shaped and formed by art into the semblance ofpersons, deities, etc. Such a point of view is necessary, and no doubt proper,for the art historian, who has to distinguish lhe 'power of images' from thepower of mere unformed things, howcver sacred and sacrifying their origins.The anthropologist is in a slightly different position, however. The 'objectswhich resemble human beings' with which the anthropologist deals, prim-áí-ily, are not portraits, effigies, idols, and so forth, but simply human beingsthemselves. Freedberg's {:mphasis on the centrality of artistic mimesis has,ânthropologically, to be set in the context in which the 'representation' of ahuman being, or indeed a deity, is most commonly undertaken, not by any kindof effigy, but by a human actor playing a role. Churches may be stuffed withimages of Christ, but the primary enactment of Christ in Christian worshipis undertaken by the priest, who plays Christ and serves as Christ's image inperforming the mass and uttering Christ's words. This is not to dismissFreedberg's question, to which I will return in the next section, but beforeleaving the subject of consecration, it is interesting to compare the two con-secration sequences we have examined, with ;1 third, in which the 'index' ofthe divinity is not a carved image at ali, but a human being.

Kumaripuja, the worship of the goddess (primarily a form of Durga) in theform of a young virgin girl, is widely disseminated in India, and is a particular[eature of the religious system of the Newars of thc Kathmandu valley, wherethe cult has been the subject of a detailed study by Michael AUen(1976). Thereare some nine 01' ten living goddesses in the Newar region, of whom the mostimportant is the one traditionally associated with the royal household. Thc virgingirl is a form ofthe fierce rayal deity, Taleju, who herselfis a form ofDurga,the violent and erotic goddess in the Hindu pantheon, the slayer of the buffalo-demon who rides on a lion and brandishes a sword. AUen's analysis concentrateson the paradox whereby a virgin, premenstrual girl comes to represem so fear-some a divinity, but what concerns us is only the mechanics ofher divinization.

Virgin-worship can, meanwhile, be distinguished from the more commontyp~ of divine (or demonic) possession which is found in India. Possession by

the deity, in this forro, is temporary and generally ecstatic; the medium goes intotrance and becomes a 'horse' for the deity, making utterances in the person ofthe deity, and 'playing' for a while (dancing, swinging on a swing, etc.; cf. Gell1978). There is no suggestion of ecstatic trance in the case of the kumari. Thisposition'is semi-permanent; once consecrated (at the age of 2 or 3), a kumariis arid remains the goddess in person until the moment of deconsecrationwhich transpires when certain 'negative signs' arrive-Ioss of milk teeth intheory, in practice, menstruation. The kumari comes from a particular caste,whoare Buddhists, traditionally attached to lhe monasteries of KathmanduwhT~h~ere disbanded by the present royal dynasty. She is, in caste terms, non-polluting but somewhat outside the Hindu hierarchical system. A candidate forkumari-hood must be old enough to walk and talk, of unblemished appearance,having lost no milk teeth, etc. Her horoscope must be auspicious, especiallywith regard to the king. Here is Allen's account of the circumstances of herinstaUation, which occur at the end of Dasain (Dassara) the festival of Durga:

At nightfall eight bufTaloes representing the demon are killed by having their throatsslit so that the blood jets high towards the shrine that contains the Taleju icon. A fewhours later at about midnight a further 54 bufTaloes and 54 goats are killed in a similarmanner. As may well be imagined, the small courtyard [ofthe Taleju temple] is by thenawash with blood ... At this point, usually about 1.00 A.M. the small Kumari-elect isbrought to the entranee. She is supposed to walk by herself, in a clockwise directionaround the raised edge until she reaches the bloody Taleju shrine. She must enter it,still maintaining a perfectly calm demeanour, and if ali is well she is then taken upstairsto a small room for the installation ceremony ... after the usual purificatory and otherpreliminary ritfs, the chief priest performs the main ceremony in which he removesfrom the girl's.body ali ofher previous life's experience so that the spirit ofTaleju mayenter a perfectly pure being. The girl sits naked in front of the priest while he purifieseach ofher sensitive body areas in tum by reciting a mantra and by touching each areawith a small bundle of such pure things as grass, tree bark and leaves. The six sensitiveparts are heI' e)I:S, throat, breasts, navel, vagina and vulva. As he removes the impur-ities the girl is said to steadily become redder and redder as the spirit of the goddessenters into heI'.

At this stage the girl is dressed and made up with Kumari hairstyle, red tiRa, thirdeye, jewellery, ele., and then sits on heI' beautifully carved wooden throne on the seatof which the priest has painted the powerful sri yantra manda/a of Taleju. She alsoholds the sword of Tale;u and it is at this point that lhe final and complete transfor-mation takes place. Ir is worth noting that though from now until heI' disqualificationsome vears larer she will be continuously regarded as Kumari, it is also believed that iris only when fully made up and sitting on heI' throne that identification is complete. Atother times, especia!ly when casually playing with friends, she is partly herself andpartly Kumari. (Allen 1976: 306-7)

To what extent can one detect a parallelism between the installation of lhekumari as a 'living icon' ofTaleju, and the installation rites of more conventional

j;~ p~il "

"il

ri~\!'{;ii.';~i(lid'n

I

I!í,

III'li'11:['

and is in fact incarnate as a human being rather than a manufactured artefact.From the point of view of the anthropology of art, as outlined in this work,there is an insensible transition between 'works of art' in artefact form andhuman beings: in terms of the positions they may occupy in the networks ofhuman social al~ency,they may be regarded as almost entirely equivalent.

FIG, 7.12/2. Kumari with paintedtltird eye. Source: AlIen ICj76

Thus I conclude this extended discussion of idolatry. I recognize, however,that the particular line I have taken has consistently resulted in the emphasisbeing placed on sociological, religious, and psychological agency at the expenseof aesthetic and artistic agency. And I am left with Freedberg's pertinent objec-fion: to traveI toa far down this road is to lose sight of art's specificity. While Ihold that where each individual work of art (index) is concerned, anthropolog-ical analysis is always going to emphasize the relational context at the expenseõf artistic or aesthetic form, the network of agent/patient relations 'in the vicin-ity' of the work of art-the same does not apply when we come to consider art-works, not 'individually' but as collectivities of artworks. So far, each index thatI have subjeclcd to analysis has been regarded as a singular entity, embeddedin a particular social contexto However, artworks are never just singular enti-ties; they are members of categories of artworks, and their significance is cru-cially affected by the relations which exist between them, as individuaIs, andother members o(the same category of artworks, and the relationships thateldst between this category and other categories of artworks within a stylisticwhole-a culturally 01' historically specific art-production system.- Artworks" in other words, ~ome in families, lineages, tribes, whole popula-tions, just likt: people. They have relations with one another as well as with thepeople who rreate and circulate them as individual objects. They marry, soto speak, and beget offspring which bear the stamp of their antecedents. Art-works are manifestations of 'culture' as a collective phenomenon, they are, likepeople, enculturated beings. So far, none of the collective issues surroundingfhe work of art have been considered. In order to broach these issues, it isnecessary to adopt a new register. Here, I can make amends for having writtenso many pages which may have seemed tangential to the study of works ofart as normally understood. I shall, for a while, desist from the terminology of'indexes' and abductions of agency, and suchlike, reverting to a more conven-tional vocabulary. Because there is one 'conventional' art-theoretical conceptwhich even the most radical anthropologist of art cannot put to one side-theconcept of style. Style, which is the harmonic principIe which unites works ofart into groups, into coIlectivities, corresponds to the anthropological theme of'culture'. Culture is style, really, just as Fernandez suggested in a deservedlyinfluential discussion (1973).

idols? As with each of our previous examples, the consecration 01' the kumariproceeds in two phases, one focusing on heI' interior and the other focusing;on her exterior. The first phase consist of the 'emptying' of the kumari-e1ectof her past life (i.e. her personhood, agency, as a mere human beir.g), whichis objectified as the impurities removed from her orifices, eyes, throat, vulva"etc. She becomes a 'hollow vessel' imo which, through the extraction of aliprevious contents, new coments may be drawn; that is, the spirit of the god-.dess Taleju. This phase, it seems to me, is the equivalem of the phase in theconsecration of the Jagganath images in which the Daitas hollow out a cavit·,in the image and place therein a (foreign) Iife-substance. But after this has bee~effected, the transformation, as Allen makes clear, is still not complete.

'fhe second phase of consecration consists of 'wrapping' the kumari in the~ress of Taleju, and painting her, as well as providing her with the goddess' sprimary ~ttribute, her sword. Allen does not make much of the fact, but themost striking visual symbol of kumari-hood, is actually the extra 'eye', the thirdeye, which is painted on the middle of her forehead. Of course, the priests donot have to paint in her pupils, as they would if she were a wooden idol-shehas very pretty eyes of her own. But they outline her existing eyes with exag-

" gerated make-up, besides adding an enormous painted eye above them (see, FTg. 7.1212). The parallel between the making-up ofthe kumari with an extra

eye and the painting-in of the eyes of conventional idols is surely rather strik-ing. In more general terms it is clear that the dressing, painting, enthronement,and provision of weapons correspond to the 'external' strategy of animationwhich I discussed earlier. Besides which, the kumari is externally animated inthe usual way by the recitatian of sacred words, and the metonymic effect 01'being seated on a magica] design of great power, the sri yantra manda/a.

111 short, there is little to differentiate the consecration of the kumari fromthe consecration of any other idol, except that the kumari can walk, and talk,

~ch individual wo~~of ~rt is the projection of certain stylistic principieswhtch form large~ u~ttte~, Just as each individual, in a kin-based society, isre~rded as a proJectton, mto the here and now, of principies of descent andalhance and exchange. The concept of style allows us for the first tt'me t. " oconcent~atc excJuslvely .o~ works. ?f art as such, and to discuss what mayappear, mdeed, to be thetr aesthettc properties. 50 let me reassure those of myreaders who are aesthetes and art-Iovers (if I have any left after ali the abuseI have sho~ered on .their ~eads)--you have reached the one chapter in thisbook you mlght posslbly enJOYreading.

Style and Culture

8.1. On the Concept o[ StyleIn this chapter, my aim is to formulate a concept of 'style' adapted to therequirements of the anthropology of visual art. In the anthropology-of-art con-text the concept of style is distinguishable from the concepts of style applic-able in Western art history and aesthetics in that the 'units' of style are not(uliualiy)inJi\'idual artists, or schools of artists, or movements, but 'cultures' or'societies'. Aetually, the units of style are conventional ethnographic isolates asrepresented for stUdy purposes by museum collections and published sourceson material culturc:. 5uch ethnographic isolates are historically bounded as wellas geographicaliy bounded; usually the collections and documents belong toa particular period. Where studies of ethnographic art in museums are con-cerned, the period in question coincides mostly with the colonial era. Thefocus is primarily on 'traditional' art forms, though it hardly needs to be saidthat during the colonial period there were startling historical developments inthe so-called 'traditional' societies of the colonial frontier, which affected theirart production in diverse ways. The problem of tradition and innovation inethnographic arts would constitute too much of a diversion to discus'i ar thisstage; besides which the subjeet has already received detailed attention fromnumerous other researchers, most notably by N. Thomas. For the purposes ofthis chapter, it will be assumed that the units of style are 'cultures' and that weare dealing with 'traditional' art as conventionally understood, setting to oneside the acknowledged problems which these assumptions admittedly raise.

The problems of historical contextualization do not materially affect what Ihave to say about style, since the description of a style may be as broad or asnarrow in scope as necessary to accommodate any given historical perspective.The question that I want to address is not historical but conceptual: what doesthe concept of 'style' contribute to the understanding of material culture?'5tyle' is a vague word of uncertain definition and many, rather disparate, uses.Finding a use for it in the context of the anthropology of material culture mightbe considered a waste of effort, were it not, in fact, so pervasive, at least as amode of classification. As it is, we are routinely accustomed to classifyingobjeets as sharing, or not sharing, stylistic attributes with one another. Butexactly what is shared (or not shared) in such instances is much harder to assess.Moreover, we are inclined to belicve that what objects with shared stylisticattributes have in common, is not just some formal, external, property, but