ART ANDAGENCY

An Anthropological Theory

Alfred Gel!

CLARENDON PRESS· OXFORD

1998

OX?'"" I./mt'mi'y I'r,-". err", CI",,,,,Jon S,'W, OxjiJrd OXl (,DP

a.lfimi ,VtW I'u,fAlh"" _4uck{,JnJ fla"K"of IJO~QI" /Jombi'l' B"",,"<_4i'"

Cakulta Cap" Tow" Dllr n .I""I,,"m Delhi "'IM'";' Hu"g Ku"! i""ff""1Kcm"'hi Kual" L"mpur Mild,", M"d"d Mrlhe""'e .Ikuio Cil)'

Nalrohi /',,,i, Si,,!tapm' '/;"p,'i Tofy" Toren/l. W""aw

,,,,d aHu,;""J i'Omp"",,, inBi-rli" Ibd,,"

1'".Ii,I,td '" Ihe U"iled Slatnkl' o.rfe'd U"it~"ity I'm, ""., A'r,,' Yo,f

(l J1lfrrd Cdl J90'Th, moral ,,:~M, o(lhr aalhor hat" "'''' """Id

All ri8ht' r""rr.',d .v" P""' '1,h,,< P"b/;''''io'' ma)' he reprud""'d.<to,d i"" "hi"t',,1 'y,l,m. IIr lrammi"d, '" anyfi"'" 0' bJ' iINy ""a,,<,

without Ih' prior pamiHi,m in ,,,,,',i"t '10xji>,d (.'"i""';I)' I'r,"",Ih'thin Ih,' UK, ,.retptio'" arc all"""d In "'N" <1~n),!air draling fiir Ih,

parp.". ofmtJ"h ur pri',"" ,Iutly. or cri'imm or ""irw, a, permimtl""J<r 'h,' Copyrithl, Di,,!"]' II"d P"tr"'" Act, '<iS8, 0' ill Ih' (a", 0/"pre~r"pl,,'e "pr.,lurtiolJ in II",wtla"", ",i'h 'h, urN" oftilt lirm'"

i,.ud hy ,I.,. Copyr(~ht Lir<nsiNl . i~",,)'. l'.nq"in", ((,",-emi"!"pr.dunion mmide ,h,,,, Irm" ""tI in other ",,""',m ",",tld hr

,,," ,,, 'h, R(~I", {kparlmml, U,j,ml !jni,'m;ly Prm.

al 'h' "ddr", a"o"~

1IIis h"uf iI ",Id ""hj", to thr ,O""ilion Ih,II;1 'hall rWl, I'J' wuyoj'lrade ur "Ihm"i",. I" lent. "-wid. hmJ ,ul,r .Ihermi" eirn"aldWilhout th,' p"bli,her", pn," ",,,,,,m In anyferm ofhinJiffg or four

orher tl,all thai i" ,,'hitk il i, p"hliIh,d ""d w"h"ut a """/,,, ronditio"includi,,! Ihi' 'o"ditio" hrill! ;mpoml ,," 'h, ",h.",qum' p,mh"""

flnrish Lilwl)' Catal"~ui"Jill l'"hli"oli,,,, Dill"

D"I" a"oil"hI,

Ubral)' ofCung"", Culal"gi"t-i"-l'ubli,'"';",, Ddt"

Grll. Alfrd.Arl ,,,,d "tme.~''- IUll'ord,,, """' u"'hn,pol".lficallh,,,ry! AIFd

G<lI.I"c/"d,']' bibliot""phi",d r,'Ji'r,'IIre,.

Arl ,>nil amhropel.gy, z, Art anil ,oeiely. I, Ti,I,.'\'72.AS6C.5 1'1')8 70,',Oj-dr21 'J7-S.fl4S

lSiL'" 0 '9-82801J-oiSBA' 0-19-82~'0l4-9(pM.)

Tyfr'''' hI' Cmph",-aj; Typ""I'" LtiI.. Ho"t K""III'ri"ud ill Crt'" 8,ilni"

.n aeid-Fe< papa hyB""hmji (Bnlh) [,td

,uid,,,mer N"rlO", So,","'t

To Simeran/IIi/lI/o;;e

FOREWORD

Nicholas Thomas

Alfred Gcll, who died in January 1997, was widely regarded as one of the mostbrilliant social anthropologists of his generation. His writing and thought wererigorously analytical, yet often also playful and provocative; he was equally deftin engaging the most general issues of social theory, and the most intricate ele­ments of rituals, practices, and artefacts. These capacities are exemplified inthis book, which may amount to the most radical rethinking of the anthropo­logy of art since that field of inquiry emerged. The book certainly combines agood deal of abstract model-making with remarkably insightful discussions ofparticular art objects and art styles.

Yet, despite it being written in a lucid and direct way, it is not necessarilyan easy book to grasp. It does need to be acknowledged here that, had theauthor lived longer, he would certainly have done further work; he indeed leftnotes toward revisions that he did not have the time to carry out. What we ha \'eis the full draft of a book, most of which was written over a period of only amonth, not an absolutely refined version. It should be added, though, thatAlfred Gell's essays and books did, for the most part, emerge well formed; hewrote with great intensity, but preferred to write when his ideas were clearlyworked out, from start to finish. The book can therefore be said to approximatean intended final form, but it does lack polishing, and there arc certainly force­ful passages that would have been qualified, points that would have added orelaborated, and sections that would have been better integrated with the whole,had Gell had the opportunity.

\Vhat the book lacks, in particular, is a preface or introduction proper, thatconcisely foreshadows its overall argument. \Vhile I hesitate to summarizeanother scholar's book, and am frankly unsure of my capacity to do justice tothe various dimensions ofa complex and involved argument, I believe that thisis whar this foreword should attempt, in order to make the arguments that fol­low more accessible, particularly to readers unfiuTliliar with Alfred Gell's otherwork. This book builds on a number of essays, and anthropologists who haveread 'The Technology of Enchantment and the Enchantment of Technology'(J992b), 'Vogel's 1\'e1: Traps as Artworks and Artworks as Traps' (1996), orGell's study of Polynesian tattooing, Wrapping ill Images (1993), will anticipatethe directions that it takcs.

The essay on 'The Technology of Enchantment', in particular, foreshadowssome of the larger arguments here. In that paper, Gell provocatively elaimedthat the anthropology of art had got vinually nowhere thus far, because it hadfailed to dissociate itself from projects of aesthetic appreciation, that are to art

\'111 Fqrtword Fur(lmml "

u

as theology is to religion. He argued that if the discipline was instead to adoptthe position analogous to that of the sociology of religion, it needed a method­ological philistinism equi\alenl 10 sociology's methodological atheism. Thisrequired diliC)\\'lling the 'an cult' 10 which anthropologists, as cultured middle­cbss intelk'Ctuals, generally subscribe. This was not, however, to adwlC'.lIe ademystifying sociologiL<l1 anal~sis that would i(lentify the role of art in sus­taining class culflll"es, or in legitimizing domin;lI1t ideologies: Gell suggesTedthat approaches of this kind r.,ikxl to enI,"::l.1,'"C with an objects Ihemselves, withtheir specificity and effiC<ICY. ",lore particularly, he was relari\'e1y uninterestedin lhe questions raised by an world instifUrions, believing instead that theanthropology of ;lr( should address lhe workings of art in general.

lie proposed that it was possible to address questions of the efficacy of theart object, without succumbing ttl the fascination and aura of those objects,by taking art as a spct:ial foml of technology, and especially by regardingart objects as de\'kes 'for securing the acquiescence of indi\'iduals in the net­work of intentionalities in which rhey arc enmeshed' (t992b: -n). For example,brilliantly involured and captivating forms such as those of Trohriand prow­boards (of lhe .\1assim region, Papua New Guinea) work a kind of psycholo­gical warfare, in a situation of competitive exchange. These boards confrontthe hos(s of exchange partners, ideally dazzle them, beguile them, and confusethem, leading them to surrender their valuables-anthropology's famous kulashells-for less than their value, T11e claim here is not reductive, however: it isnot sugges(ed that in some sense the object kJ' itselfdoes this, or would do i(,independently ofa field ofexpectations and unders[andings, \\ hich in [his caseenvelope the arl.efac[ with magical prowess, which is known to have enteredinto its making. Tcchnology is enchanting beC:lUse it is enchanted, 1x.'CJ.useit is the OUlcome of some process of b;lrely comprehensible \'inuosity, thatexemplifies an ideal of magical dficacy that people stTuggle to realize in mherdomains.

There was a minor inconsistenC) in the 1992 aniele, in the sense that itseemed to be assumed that the anthropology of an remained the study of'primitive' an (Gell rejected Ihe euphemistic term 'non-Western' on thegrounds that this includcd high Orienml art and other traditions, which dcarlypossessed an enlirely different social local ion 10 the eanoniealtribal art forms).I-Io\\e\"er. Ihe examples he proceeded 10 usc, in pointing to the 'halo effect oftechnical difficulty' and other aspects of the art objects, included I he paintingsof the American illusionist J. F. Peto, and Picasso. 'fhe implication that his theorymight in fact be a 1heory orthe workings of alt art. rather than that supposedl\'eharaeteri~tieof particular popularions. is a premiss of Ihe present book, .

The first ehaplers, amou", to a. dramatic elaboration of the arguments of the1992 l..'Ssa~. Gell beg-ms by defeITIng 10 the desirabilin· in broad'" e••I, I d

"

, . , ... ura anpo ItK"31 Icrms, of ad..now1c~lging the distinctness of non-Western ,Ies[hcticsystems. but asserts that thts cannot eonstitu1e' .", <, tl I' I' I. "n nopo oglca tleor)',

on the grounds that amhropologil.."31 theories :lrc l-'Ssentially concerned withsocial relations, over the l'ime-fnlllle of hiographies. As he acknowledges, thisdefinitional orientation may be contentious, but il arguably pro\'ides a product­i\'e departure poim for rhis panicular inquiry. There are two linked argumentsfor a shift away from cross-culrurJI al-"Siheties. The first is that man~ canon­ical pieces of tribal art, such as rhe Asmat shields of south-west New Guinea,arc plainly not intended to elicit 'aesthetic' appreciation in the conventionalsense-they rarher had a part ru play in the deadly ps~ehol{lgical warfare ofheadhunting, that was so fundamental to Asmat sociality before pacification.The second is a <.<Itegorical rejection of the linguistic analogic<; that have bl-'Cnmobilized by so m:my semiotic and symbolic theories of an. And rhis is per­haps the sense in which this book is most radical. For many schohlrs, andind<.'Cd in much eommon-sense thinking about an, it is axiomatic that art is amaner of meaning and communication. This book suggests that it is insteadabou( doing.

'Doing' is theorized as agepC)', as a process ill\"olving indexes and eHeets; rheanthropology of art is constructed as a theory of agency, or of the mediation ofagency by indexes, understood simply as material entities which mOlivateinferences, responses or interpretations. Indexes stand in a \ ariety of relations1O prolO~'pcs, artis(s, and n:cipients. ProlOtypes arc the things [hat indicesrna} represent or stand for, such as the person depicted in a portrair-thoughthings may be 'represented' non-mimetically, and non-visually. Recipients arcthose whom indexes are raken 1O effect, or who may, in some CJ.5eS, be eflccti\'elhemseh·es via the index (a view ofa country est'ale commissioned by the land­owner may be a \'ehicle of the recipient's self-celebrating agency, more thanthat of the artisr. Artists are those who are considered (l) be immedia(elv GHlS­

ally responsible for the existence and characteristics of index, but as ~e havejust noted, 1hey may be vehicles of the agency ofothers, !lot the self-subsistent,crcati\-e agems of Western commonsense idC'.lS and art-world theory, In thisrespect. it is worth noting thar despite the notable differences bel\\een the styleand orientation of this book, and the Mclanesianist deconstructionism ofMarilyn Stl'athern (1988; sec also Wagner 1992), Gell could be seen to fullyembrace Strathern's notion of the 'partible' or 'distribuled' person, and indeedto make explicit the ways in which it follows from this (:oncepr that actions andtheir effects are similarly not discrete expressions of individual will, but rather~he outmmes of mediated practices in which agents and patienls are impli<'"3ledIII :o~lplex ways. On the one hand the agency of the artist is rarcly sclf­Sl1~elent; on the other the inde.~ is not simply a 'product' or end-point of:tellon, bUI rather a distributed ex(ension of an agent. The chilling example ofone of Pol Pot's soldiers, who distributcs clements of his o\\n efficacy in thefonn of landmines. is one ofthe many unexpected, )et apt instances that giveswhat \\ould otherwise be an intractable, abstract c.\position of these terms,SOme concreteness.

, For(U'ord

The theory f(,'(.'Cin:s morc sustained exemplification in Chapters 6 and 7.which address forms of'dccorati,'c' and 'representational' art respecti,'e1y. Thefirst discus,'OCS apor[opaic patterns, in\'olured designs intended (0 corr-.mecand ward ofT dangerous spirits; \\;th examples such as the Asmat shield, theseperhaps maniresl most ob\'iously onc of the book's larger theses, namely thatart objl.-'Ct.<; mediate a technology to :lI;:hic\'c certain ends, notably W enmeshpatients in rcl:ltions and intentionaliries sough I or pn:scribed by agenrs. Lestthis appear a reductive apprOolch to art, one thar lakes objects essentially asvehicles of strategies, it is importanllO emphasize thar the formal complexity,and indeed the technical virtuosity, exhibited in works of art is not incidentalLa the argument but absolutely central to it. It is crucial TO the theory, in fact,

./that indexes display 'a certain cognitive indecipherability', that they tant:llize,(J they frustrate the viewer unable to recognize at ollce 'wholes and parts, conti­

nuity and discontinuity, synchrony and succession'. Enn though this hookengages in little sustained cognitive theorizing, it is nmable at thi!> point andelsewhere that cognitive observarions animate Gcll's argumem, lO a degree thathas become unusual in anthropology.

The long chapter which follows ranges widely Mer idolatry, sorcery, ritual,and personhood, and incidemally displays Gell's grasp of a bewildering rangeof south Asian and Polynesian source material, but is fully consistent with theclaims of the previous sections. Idols are indeed of special relevance for rhebook, because they stand for an agent or patient (in the case of sorcery), for per­sons or deities, in manifest and powerful ways. They arc indices that may be:mimated in a variety of ways, that enable transactions in lethal effect, fertility,auspiciousness, and the like. The particular forms of agency and intention atissue here, and the process of consecration, arc explored in detail. 'fhe largerpoint is that there arc multiple implications ofagency in objects, 'an insepara­ble transition' between thcm and actual human agents. Once appreciatedas indexes of agency, iconic objct.'tS in particular can occupy positions in thenetworks of human social agency that are almost equi\·alcnt to the positions ofhumans themseh·es.

Cp to this point, Gell's theorizing and exemplification h:l\·e focused uponthe work of particular objects or indexes in particular actions, on specificprocesses rather than entire repertoires of art\,orks. He concludes Chapter 7by acknowledging: thal there are many vital respects in which artworks do notappear as singular entities, but rathel' as ensembles. The remainder of the buokappears to take ,I sharp turn :iway from the paradigm of the agent and in<lexthai h:is received such concerted :lttention thus far. It tackles the question offamili:ll relations among artworks, and seems 10 shift back to cOIwentionalground, in engaging \\·ith the concept of style. Yet this discussion, which pro­ceeds via a rich formal analysis of the extraordinary corpus of ?\larquC&1n artdocumented by the German ethnologist Karl ,·on den Steinen, is in the endcOllsistCnl wilh what comes before.

Gell is alf.lin concerned to a\·oid linguislic models such as 'a grammar ofst\'le' and instead seeks rather to identify axes of coherence through a stricti}fo~al analysi.<; of generati\·e relations among motifs. The bulk of the chaplerconsists of a richly \'isual analysis of these relatiuns. The point that Janusfigures (which arc almost pervasi\'c in Oceanic art) indexed invulnerability hadalread,' bcen made in IVmppillg in Images; it was not simpl} that the figurecould ~(:e in all directions, hut that the face was itself an expression of power,:ind, in sculptural form, was canonicllly the face ofa deity rather than a human.One of the central claims of '''rapping ill images was that eastern Polynesian tat­tooing W:iS a technology that reinforced the body, and in the highly competi­tive, unstable, and viulene societies of the Iv1arquesas, iT is not unexpcctcd tofind that tattooing entailed the multiplication of the hody's f;ltcs. These themesarc highly salient to Gell's discussion of Marquesan forms such as the famousIt'll clubs, described here as 'the ultimate double--double tiki', hut the chaptergoes well beyond the earlier discussions of the arts of empowerment in thesesocieties. The real objL't.1:, in [his {:asc, is the diagnosis of the formal principlesthat glve MarquL'San art irs singularity, and these arc identified, not at the le\'clof appearance, but through the types of transformations thal 1in1: Marquesanartworks. d-

At the most abstrJct le,·c1, the principles that go\·em these transformationscan be connected to the cultural milieu. Gell suggt:st.-; that the most basicprinciple w be detected in the Marquesan corpus is a principle of 'least differ­ence': 'the forms taken by motifs and figures are the ones involving the leastmodification of neighbouring motifs consistem with the establishment of adistinction between them.' This [rend can in turn, he claims, be connectedwith the most basic feature of identity-formation in Marquesan society, whichwas characterized by acute status competition; this was nllt simply a matter ofpolitical jockeying, but rather a ritually &1turated process of inter-individualCOntact and commensality, Personal intL-grity was continually thrL"'atened bydispersal and de--differentiation; many ~iarqucsan artefacts amounted, indi­\idually, to deyiet:s that wrapped the body and protected particular orifices,or the body as a whole in situations of crisis; in the ensemble as a whole, tbeprinciple of least difference resonated with a preoccupation with a continuallyprejudiced effon of differentiation, of differentiation in the midst of dissolu­tion, 'There was an elective affiniry between a fIlodw r,peralldi in the artefaClUaldo~a~n, which generated motifs from other motifs by intetpolating minuscule\,;lrIatlOns, and a modus operal/{ii in the social realm which created "difTerences"arbitrarily against a background uf fusional sameness.'

It is \\·orth underlining the distinction between this dfort and that of AlbnH~nson, ~\'hich Gell finds, in an opening section ofthe chapter, to be worthy bUImlSConccl\·ed. \\'hercas Hanson ancmptcd to identify one-to-one Lurrespondenec<:between formal properties in ,\Iaori :lrt (sueh as disrupted symmetry) andProperties of ~iaori culture (compctiti,·e reciprocity), Gell points OUI that lhe

XII FormN,JrJ Forrword xiii

stylistic clements that arc singlcd out are uniyersal, or at least (;ommonlyencouOlered, and cannot therefoTC be determined by singular features ofMaoriculture. i\lthough the 'electh"c affinity' that he ~ks lO identify between Mar­quesan style and culture could be seen to be similar to the relation of recapit­ulation th:lt Hanson postulated between Maori aesthetic form and culture,for Gell the affinities will emerge not at the level of characteristic rclations inparricu1:Jr bodies of m:nerial but at that of 'rclations bet ween relations'; l!!.- thelevel, in other words, of meta-properties that demonstrably render that stylepeculiar to itself.

The final chapter makes a further, equally ambitious step, on to ground thathas often been unsatisfacwrily tra\'crsed, the problem that has been co",'en­tionall} posed in terms of "hat collective counterparts individual minds andoonsciousnesses possess. Gel!'s appmlch to the issue may be fresh and reward­ing precisely because it docs nOl srarl from the usual departure points, butrather builds on several preceding arguments-'inner' and 'outer', internaland external, have already been shown to be relatively r:nher than absolutelycontrasted. Inspired by Peer Gym's onion, by Strathern's fractal conception ofpersonhood.... and by the eXLraordinary exemplification of fractal and distributedpersonhood in Polynesian and especially Marquesan arr, Gell evokes {he notionof a 'distributcd mind' through an argument thar 'lhe structures of art hiSTOrydemonstrate an cxternalized and collectivized cognitivc process.' The famous!o.Ialang-.1.Il of !'\ew Ireland and the Kula transactions of the ';\hssim region ofPapua New Guinea arc ill\oked to advance this argument, demonstrating, withthe support of the work of Kancy ~Ilunn in particular, that the Kula operator'is a spatin-temporally extcnded person'. At this point two of the book's keythemes, that of the distributed mind, and that of efficacious agency-uponwhich so much emphasis is placed in the opening sections of the book-aredrawn together. Eflicacy is founded 011 a comprehensive internal model of theoutside field. One become." a great Kula operator, in other words, by modellinga working simulacrum, a dynamic space-time map, of the play and hiSTOryof Kula in the world. Internal mentJI process and external rT:lnSJ,etions inobjectified personhood are (ideally) fused. Mind, therefore, em exiSt object­iwly as well as subjectively, as a pattern of transactable objects.

Gell does not conclude wirh this large claim, but pn:K:<..'eds to ,·indicate theconcept of the distributed mind through lhe more f,'miliar instance oflhe indi­vidu.ll (c.monically Western) artist's work, LUrlling also to eng'Jge with qm,"S­tions of cominuity over time, and foreshadowing the concluding discussion ofquestions of tradition. His key Lerms here arc 'protcntion' and 'retention',which ad,·en to the wa),s artworks at once anticipate future works and harkback 10 others. His key example is the a:uvre of Duchamp, and particularlythe very striking nOlion of'rhe nctwork of sTOppages' which inspired nor onlyGell's understanding of the issuc, bur rhe diagrammatic fonn in \\ hich he pre­scnts ;1. The final st."Ction of the boo!.. re'·cns to the colk·eti\·e register, acb'Uing

thar a similar pattern of protentions and retentions can be identified in thchistory of ~Iaori meeting houses, therefore understanding Ihis historical cor­pus as 'a dis{ributl.-d. objecr structurally isomorphous to t.'Onsciousm."S." as atemporal process'. There are many incidental al..l.:omplishments of this discus­sion, such as the demonstration of the extent to which 'fractJl personhood',a concept fashioned and largely isolated within J'oIlelanesianist anthropology,possesses great saliencc beyond it.

This is a demanding book. The rallgc of Lhe examples that are discussed indetail is quite hreath-taking, as is the enscmble of big conceptual questionsthat arc tackled. It will incvitably be contentious: many :l11thropologists of arthave exhibited great virtuosity in semiotic interpretations, and will no doubtrcmain unpersuaded that an approach which eschews linguistic analogies andconcepts can represent an advance on thcir own. Regional specialists, such asPolvnesianists, may be taken aback by the unexpected character of Gell's wayof s~eing. Yet the fertiliry of his provocation cannot be questioned. His specificclaims concerning Oceanic and other materials give specialists a chance 10

move beyond the interpretations, too often bland interpretations, that havcassumed the status of received wisdom; while the unprecedented effort totheorize fundamental questions of pcrwnhood and cognition from the vantagepoint of a theory of art may be as destabilizing and suggesri\'c for the formcras for the latter. Fril..'Ilds and colleagues remain painfully conscious of our lostopportunities to debate the issues further with Alfred in person; yet he hasleft us with a distributed clement of his own personhood, an index of his owncreati,"e virtuosity, a gift.

ACKNOWLEDGEMENTS

This book chronicles a momentous and sad lime for the many people wholoved and admired Alfred, and a clt':lclysmic one for me and my son Rohan,and for Alfred's parent's. Alfred began the final chapler directly upon rccci\·jngconflnn:llilln of an incurable condition; in the weeks that followed he devotedconsiderable time towards giving dear instructions concerning its posthumouspublication. He had abSolute confidence in Pel,cr ,\1omrchiloff of OxfordUni\crsiry Press, with whom he had previously liaised. It gives me pleasure to

record lhat this confidence was amply justified in the months followingAlfred's dL-ath just three days after the posting of rhe manuscript. With theexception of two illustr-Jtions pertaining to Chapler S the book appc3rs in theform in which Alfred len il. h amounts to a first drnfr written during an intensethree-week period over rhe Easter v;\CUion of 1996. Alfred intended to drawwith his own hand many of the photographic iI1ustr:nions and had begun to

make notes towards rhe revision of the manuscript shortly before he was takenill. Dlessed with implacable contentment and gtxJd humour, he was pleased andgrntcful for being gramed time enough in which to squeeze completion of abook on art-the subjel1 that truly grippt-d and ddighted him-~mdwas satisfiedwith the final pmduet.

It musl be made clear that the people listed herein were collaborators whosecommitment to ensuring the book's smooth publication, while different innature from mine, arose OUI of a direct personal invoh'ement with Alfred.Stephen and Christine Hugh-Jones had lx'Cn at school with Alfred. Stephendevoted many hours of his time to scanning the illustrations so [hat themanuscript could be sent to the publishers while Alfred was alive. He com­bined [his with keeping a watchful, solicitous cye over Rohan and, in commonwith Christine, with providing 1\lfred with an openly affectionate companion­ship during his last wceks. Alfred, nCYer at a loss to respond nrbally, wassignificantly silent wher. it came to his notice that Stcphen was purring in somuch work. Nicholas Thomas and Alfred had enjoycd a mutually stimulatingintellectual partnership that evoh'ed illlo a dose friendship over many years: itwould be fair to say that he has taken as active a role as mine in all matters relat­ing to the publication pmc(:ss. Chris Pinney g;l\ e conSl;lllt moral and intellec­tual support and Michael O'Hanlon arranged fur permissions in relation to theillustrations held at the Brilish Museum. Other friends of Alfred's :md mine,namclv Don Gardner, Carrie Humphrey, Eric llirsch, .Marilyn Slrathern,Suz;lll~e Kiiehler, :lnd Howard Murphy, attended to various queries that cameup in the (,.'Ourse of publication. Sus:m Gell, Alfred's mother, and Trudi Binnsredrew some of the illustrations from rough s"etches of Alfred's.

A(J:!IolPledgelll~/ils

On grounds different from those fur the people mentioned above, I wouldlike to thank those who responded to my requests for help: Anila llurle~" Doni\lanning, Paul Caldwell, Aidan Baker, Alison Dccprose, and i\1adhu Khanna.The following institutions generously wai\·cd normal reproduction charges:The British ;\luseum, London; .\lusCc de I'Hommc, Paris; Haddon .\'luseum,Cambridge; :"Japrstek l\'luscum of Asian, Afri<.-an and American Culture,Prague; RijL:smuseum H)Or Volkenkunde, Lciden; and Germanisches National­museum, Nuremberg. And finally, I wish to record a personal debt lowe toRobert Ritter of Oxford Universily Press. As lhe person in charge of produ­cing a complicated book, Robert has impressed me by his constant sensitivity,patience, and warmth towards me.

SL'IHlAN Gf.I.L

CON'T'ENTS

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

I.l. em there be an Anthropological Theory of Visual Art?

1.2. The 1\l"t Object :;

1.3. Art Sociology 7I.4. 1'hc Silhouette of an Anthropological Theory 10

2. The TheOlY oIthe Art Nexus 12

2.1. Constructing a Theory: Terms and Relations J2

2.2. The Index 13

2.3. Abduction 142.4-. The Social Agent I fi

2-5- 'Things' as Social Agents 17

2.S.I. Paradox Elimination 19

2.5.2. Agents and Patients 21

2.6. The Artist 23

2.7. The Recipient 24-

2.8. The Prototype 25

2.9. Summary 26

3. The Art l\lexlIs and tlte Index 28

J.I. 'fhe Table of i\gentfPatient Relations between Four Basic Terms 28,3.2. Index-A~ Artist-P 28

3·3- Index-A -------7 Rccipicnt-P 313+ Index-A -------7 Prototypc-P 32

3·5· Artist-A -------7lndcx-P 33

3.6. Recipient-A -------7 lndex-P 33

3·7· Prototype-A -------7 lntlex-P 3S

3·8. The Centrality of rhe Index 35

3.8. J. The Logic of 'Primary' and 'Sccondary' Agcnts

and Patienrs 36

3.9. The 'Illegitimate' Expressions 38

x"iii CimtenLS

3.10. Index-A -----lo Index-P3.11. Artist-A -----lo Arrist-I)

3.12. Recipient-A -----lo Recipicnt-I'3.13. Prototype-A -----lo ProWtypc-P

4- The II/volution ofth, Inde.\' in the Art Nexus4-.1. Hierarchical Emlx:dding of AgcntfPatiem Rclations-l.2. The Effect of Substitutions

4-.3. Trcc-Structures

........ Some more Complex Tree-Slructures: The Nail Fetish

5. The Origination ofthe Int/ex5.1. Agency

5.2. C:l.ptivation

6. The Critique vIthe Jnt/e.t'6.1. On Decorative Art6.2. Attachment

6.3. Decor:Hive Pattern = Index-A -----lo Index-P

6·4· SYlllmetry and the Appearance of Animation6.5. Complex Patterns

6.6. Complex Panerns as 'Unfinished Business'6.7. Taste and Tackiness

6.8. The Apotropaic Pauern

6.9. Ko/am

6.10. Kola"" Tattoo, and the Cretan Ma7-C

6.11. Sand-Drawings of Malakula and the Land of the Dead6.12. Orawing and Dancing

7. The DistribuMI Penon7. I. Mimesis and Sorcery7.2. The Mimetic Faculty

7.3. Volt Sorcery7+ The t:picurean Theory of'Flyillg Simulacra' as Parts of the Body7.5. From Sorcery to the Cult of Images7.6. Decortication and the Exchange oflndexes: Tahirian To'li

7.7. Darshal/: Witnessing as Agency7.8. Animism and Anthropomorphism

7.8.1. Stocks and Swm..'S7.9. External and Internal Conceptions or Agene)

COli/aliI XIX

.' 7. 10. The Animation of Idols: The Externalist Strategy I,J,J

..5 7. 11 . Concentric Idols and Fractal Personhood '37

..7 7. 12. The Rites of Con$l.'(,;ration '43-18 7. 13. Conclusion: From the Indh-idual [0 Lhe Collecti\'e 153

-, 8 Style flnd Culture 155,51 8.1. On the Concept of Style 155-, 8.2. Ilanson on Style and Culture 159,-53 8·3· St)'le and Cognith'c Saliency 162

59 8.... Formal Analysis and the Linguistic Chimera ,63

66 8·5· Synecdoche: Axes of Coherence in Stylistic Unities 1658.6. The Marquesan Corpus ,68

6(,

68 8·7· The Table of EllUl Motifs '7'8.8. From HI/pI' Vehill( to Va; 0 Kella 174

73 8·9· From Hllpe Vehil/e to KakefKea '76

73 8.10. From the Ell/a to the race Motif (Matfl HIiIlIa) ,So

7' 8. [I. From Mala /-Ioala to Ipu: Additional Face Motifs 182

76 8. [2. Some Additional Motif-Transformation ,S6

77 8.13· Figure~Ground Reversal (Tortoiseshell Diadems) ,87

79 8.1-l. From Two to Three Dimensions 19°80 8.15· Backlcssm.'ss and Split Represcm:uion: Shell Crowns '928, 8.16. Janus-Faced Images: Scale Transformations 196

83 8.17· .Multiplication, Transposition, and Proboscis-Formation:

S.. Fan Ilandh:s 202ll(, 8.18. The Same Continued: Stilt-Steps 205

9° 8.19· U"I: The Uhim;l.le Double-Double-Double Ti/n 20<)

9.. 8.20. Fusion: The Narrati\e Art of Ear-Ornaments 2ll

8.21. Conclusion: Coherence in Marquesan Art and Social Relations 21 596

96 9· Conclusion: The Ex/ended Mind 221

99 9·1. Distributed Objects 221

'02 9.2. Malangan 223

[°4 9·3· Gawan Kula >28106 9·4· The Artist's (EI/vre as a Distributed Object 23 2

'°9 9·5· The (f'lIvre of i\hrcel Duchamp 242116 9·6. The Maori i\leeting I louse '5'121 Bibliography 259l22

126 Jndex 265

Lis! o( Figures '"6.J If I Diagram traced by the Ambrym natives to explain their

kinship system 9'LIST OF FIGURES 6.11/2 Lcvi-StnlUss's diagram to explain Aranda kinship in

The Savagr Mind "'.Iron/is. Asmat shield XXIV 6. If/3 l'vlalakulan sand-drawings 9'3.8 .1/ 1 The index as the pi\'ot of the an nexus 38 6.rr/4 Nahal: 'the Path' 93J.1O/1 The self-made index 4' 6.I2JI Performance art complementing graphic art: the j'vlalakulan

3. 10/2 The black disc looks as if it 'wanted' to return to the centre dancer's path 94of the square 43 6.12/2 'The Hawk flying from place to place' 94

J.IO/J Causality made visible: detail from Plu/o ({nd Persephone by ].3/' Volt sorcery and the paradoxical agency of representatiun 10JBernini . 44

7.5/1 The circularity of productive exchanges between index4. 2/1 The prototype as agent: Samuel Johnson by Reynolds 53 (mauri) and prototype (hall) 1094-3/ 1 ·rhe multiple levels of agency within the index 54 7.6/, Ti'i in its benign form as a volcanic gTowth stone "04·3/' The hierarchical embeddedness of ~lgent-patientrelations 55 7.6/, To '0 with Oro The god hidden from view and wrapped4.4/1 Nail fetish figures from the Congo region of Africa 60 in bindings '"H/' The fetish as an index of cumulative agency and as the visible

7· GIJ Sorcery li'i with hollow back for feathers/exuviae '"knot tying together an inyisible skein of spatio-temporalrelations 6, 7.7/ 1 Santoshi Ma I HJ

H/3 Illrs Pankhursl by l'vlary Richardson: the Rokeby VenllS by 7.7/2 Ocular exchange as the medium for intersubjectiyi Tics "0

Velazquez slashed by Mary Richardson, 1914 63 7. 11 /1 The fractal god: A'a [rom Rurutu IJ8

H/4 The shared biographical spaces of persons and images 65 7. 11 /2 Genealogical personhood objectified: a staff god from the

5. 2/1 Trobriand canoe prow~board 7°Cook lslands IJ8

6.2/1 Iatmullime containers 757· TI /3 'Vierge ouvrante' '4'

6·3/' The hierarchical structure of the decorative index 777· TI /4 An Indian parallel to 'viergc ouvrantc' 143

6·4/' The basic motions of pattern formation 78 7. 12/1 'World-seed': sa/agrama '46

(q!2 The linear repetition of a single motif: simple p~mern 78 7. 12/2 Kaman' with painted third eye 152

6·5/' The multidimensional orientation of a single motif: 8·7/' Efaa '7 'complex pattern 80 8.7/2 Plane rotarion in the formation of Steinen's two chlsses ofe/ua '7'

6·5/' The pattern as mind trap: glide reflection and 8·7/3 Reflection and rotation in the formation of c/ua L, M, and N

figure-ground reversal 80 in Fig. 8.7/1 ,-,I.'6.8/ ' Celtic knotwork: apOlropaic pattern 84 8·7/4 Geometric orientation affecting the change of mutifs:

6·9/' Ko/am threshold design: the pattern as topological sn:lre 8- from clua to the female tattooing tortoise motif kea 174,6.9/2 The constitutive element of the ko/am 86 8.8/1 ElIw 0 in Fig. 8'7/1 sliding by degrees to become the

6.10/1 Comparison of threshold, tattoo, and coin designs fromhope uhine class of motifs '74

Layard 1937 87 8.8/, From sining elua to hope vehillC 175

6.1O{2 The serial COllllection of points in the formation of 8.8/3 From hope vehillf to vai 0 Ketlil '76the Cretan labyrinth design 89 8'9/1 Coordinate transformation of a hope vehinc motif resulting

6.10/3 'rhe Roman maze at Pula 89 in /.take (type I) '76

XXII Li1t ofFi/(urn LUI of F~l!urrs xxiii

fLO/2 KaKe from hope 'veltille: dupliC:ltion, rotation, and reflection [77 8.,6/6 The rules of body proportions in Marqm:san art '0'8·913 Coordinate transformation with rOlation: from sitting cilia S.lil r Marqucsan :lrtisric pl'Olifer'ltion: fan handles divided into

to kak~ (type II) 177 multiple scgmcms ill differcnr orientations 2038·9/, The chicfL'S!>'s girdle with type til kak~ '78 8.17/2 Double-anchor fan handles with morphologicall~ adapted head '0+8·915 Ken (tortoise) female tatwo motifs recorded by Steinen 179 8·'7/3 Fly-whisk handle from the Austral Islands with similar proboscis

8·916 A tortoise of tortoises: kea motifs on tortoise-shaped terrinc 179 fa<.:e 2°4

8·9/7 'Realistic'lua IOrroisc tanoo recorded by Langsdorff 179 8.18/1 A simple form of stilt-step with rudimenrary face at base 2058.10/1 Silting dua hypostasized to mala/wala ,8, 8.,S/, From rudimentary face to complete second figure: stilt-step '068. rO/2 Malll Iwaftj in male tattoos recorded by Steinen ,8, 8. ,8/3 T ransformarion of nose/mouth to proboscis 206

8.10/3 The mala hoata latWo indicated lends cvidently towards 8.,SI+ Stilt-step with multiple figures and 'mooning' bUllocks 207the ~/lla motif ,8, 8.18/5 Slih-step with rotated duplicate figure '07

8.10/4 from mata hoaia to the Vall rock morifby the addition ofspirals ,8, 8.,8/6 Maximally proliferated srilt-step with eighr figures 2088.[1/' From double mala IlOa(a to ipu ,8, 8.19/r V'u warrior's club with protective face proliferalion 2098.11/2 Tapa mask with ipu 'eyes' ,8, 8·'91' The figurative model for U'u dii>playing subsidiary figure

8.1 rl3 'Hand f.'tce' dcsign I 18] on the hack '"8.11/4 'I-land face' design 2 ,83 8·'9/3 V'II exemplifying Stcinen's model wjlh subsidiary head

8.1 r/s 'Coiled shellfish' poriri as a face motif (84 protccting the 'back of the back' of the club '"8.IIJ6 LangsdortT'i> 1813 cngraving of young warrior with eye and 8.20/1 Narrative art: 'Pahuatiti's daughters on their swing' depicted

face designs ,85 on ear-plugs ,,,8.n/I Rosette mOlifs ,86 8.20/2 Ear~plug "-irh swing narrali"c moving IOwards the

8.12{2 Hieran:hization: 1/illO 'teeth' motifs ,87t'llia/mlJIII kO(l!a morif "3

8.12/3 Auialill band containing alternating revcrsed ewa motif I8.20/3 Ear-plugs depicting the obstetric myth dii>playing a fusion

in Fig. 8.7/1 ,87of bodies 214

8.13/1 Figurc.:-ground reversal and hierarchization: tortoiseshell 9. 1/' Malang-an can'ing 22,!-

incised disc ,88 HI' The artist's «livre as a distributed object '358.13/2 From cyes to lizard: Steinen's cvolutionary derivation from 9-4/2 A version of Husserl's diagram of time-consciousness

end-plate discs ,89 from Gell, The Al1lhropology ofTim~ '398.1411 From three dimensions to two in Western art: the back view 9·5/1 Tht Largr ClaH or Th~ BrIde Jlrippd bar~ b), h~r Ba,hdoTS

of the image as the back view of the prototype '9 'rom by Marcel Duchamp 245

8.15/1 From two dimensions to three in ;\1arquesan art: providing9.5/2 nit' Ne/mork ofS/oppages by Marcel Duchamp 246

the image with a back '9' 9·5/3 TIle YOl/llg Ma" am! Girl in Spril1g by Marcel Duchamp (r9 11 ) ,+8

8.15/2 Fmm Levi-Strauss 196]: split representation and the 9·61' The Maori meeting house 3S a collective index of

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

8.16/1 SlOllC terrace of ceremonial platform '979.6/, Kowhaiwhai patterns

254

8.16/2 A'fahi-(aull-pe/N: femalc di\'inity giving birth9.6/3 The Maori mccting house as an object distributed in space

'97 and time8.1613 Wooden poStS carved in [\\'()-dimensional form '98 255

S.16/4 The double liki fonn in bone hair ornmncnts withTABLEhaJ,'eI/J:reuz motif on back, 2nd row '00

8.1615 The Iwkrnkrellz motif: a \arianl of twa via hope "/}('hi,,(' '00 The Art Nexus '9

Asm"l shield. SOllrce: Rijk ,vlu~cum VOIlT Vlllkl'llkunJc (R..\IV 1854---146)

1

The Problem Defined: The Need for anAnthropology of Art

1.1. Can there be (/11 Anthropological Thew)! oj Visual Art?

An 'anthropological theory of visual art' probably suggests a theory dealingwith the art production in the colonial and post-colonial socicties anthropo­logists typically study, plus thc so-called 'Primitive' art-now usually called'ethnographic' art-in llluseum collections. The 'anthropological theory ofart'equals the 'theory of art' applicd to 'anthropological' art. But this is not whatI have in mind. The art of the colonial and post-colonial margins, inasmuchas it is 'art', can be approached \'ia any, or all, of the existing 'theories ofart', in so far as these approaches are useful ones. Critics, philosophers, andaesthelicians have been busy [or a long time; 'theories of art' constitute a vast'and well-established field. Those whose profcssion it is to describe and under­stand the art of Picasso and Hrancusi can write about masks from Africaas 'art', and indeed need to do so because of the vcry salient art-historical rela­tionships hetween the art of Africa and twentieth-century \\-'estern art'. Thereis no sense in developing one 'theory of art' for our own ~lrt, and another,distinctively different theory, for the art of those cultures who happened,once upon a time, to fall under the sway of colonialism. If Western (aesthetic)theories ofart apply to 'our' art, then they apply to cverybody's art, and shouldbe so applied.

Sally Price (H)8lj) has rightly complained about the essentialization andconcomitant ghenoizarion of so-caned 'Primitive' art. She argues that this artdeserves to be evaluated by Western spectators according to the same criticalstandards we apply to our own art. Art from non-\Vestern cultures is not essen­~iany differem from our own, in that it is produced by individual, talented,Imaginative artists, who ought to be accorded the same degree of recognitionas Western artists, rather than being viewed either as 'instinctive' childrenof nature, spontaneously expressing their primitive urges, or, alternativc1v, asslavish exponents of some rigid 'tribal' style. l.ike other contemporary writers(:n. the s~biect of ethnographic arts (Coote 1992, 1996; _l\lorphy 1994, 1996),I nce beheves that each culture has a culture-specific aesthetic, and the task ofthe anthropology of art is to define the characteristics of each culture's inher­ent aesthetic, so that the aesthetic contributions of particular nOll-\Vestern

2 HI' Prt!h/~1J/ lliji"tJ fhr Prohfrm lkfittrJ .1

artists can be evaluated corn:ctly, that is, ill relation to their culturally specificaesthetic intentions. Here is her credo:

The crux of the problem, as I understand ii, is Ihat the appreciation of Primiti\'c Arthas nc-Jr1~' a1wars txen phmscd in terms of a r.IU<lciuus choice: one option is 10 lei theaesthetically discriminating eye be our guide on the basis of SOIllt: undd"incd conceptof universal beauty. The uther is to bury ourselves in 'tribal lore' to dis(.:uvcr the ulili~

tarian or ritual funclion of the: objects in question. These IWO routes arc generallyviewed as compelilive and incomp:uible ... I \\ould propose the possibility of a thirdcollCcptuali.....ation Ihal sits !>Omcwhcrc bet\\een the IWO extremes ... lr Tctjuires theacccprancc of £\\0 tenets that do not as yet enjoy widespread aca:ptancc among cdu­cued members ofWeslern socielies.

-One lenel is that the 'eye' of en:n lhe mosl natur:llly gifled (.'Qnnoisseur is notnaked, but \·jews art through lhe lcnsc of a Western cultur.11 education.

The second i~ that many Primiti\'es (including both artists and critics) arc alsoenduwed with a discriminating 'eye'-simiJarly fined with an optical device thatrcflcct~ their own cullural educalion.

In Ihe framework ofrhcsc two leneL~, amhropologiC'JI oonrextuali1,.:ation reprcscnt~,

not it ledious elaboration of uotie custOms th:a! oompctc:s with true 'aesthetic experi­ence,' but rather a llll'11ns to expand the acsthetic experience be~-ond our own narrowlyculturc-bound line of \'ision. Ha,ing acceplcd worl.:..~ of Primiti\'t: Art as wonhy ofrepresentation alongside the works of our own societics' mOSI distinb'Uished artists ..our next task is to acknowledge the existence ,md legitimacy of the aesthetic frameworkswithin which they were produced. (Price 19119: 92-3)

This view is perfecLly consistent wilh the close relationship betm.:cn arthistory and the theory of art in the West There is an ob"ious analogy berween'cuhure-specifie aesthetics' and 'period-specific aesthetics'. f\rt theorists suchas Baxendall (1972) ha'·e shown that the rcception of the art of particular per­iods in Ihe history ofWesrern an was dependcnt on how the art was 'seen' atthe time, and that 'ways of seeing' change over time. To appreciate the art ofa particular period we should try to recapture the 'way of seeing' which artistsof the period implicitly assumed their public would bring to lheir work. Oneof Ihe art historian's tasks is to assisr in this process by adducing the hisroriealcome.xt. The anthropology of art, one mighl quite n.:asonably conclude, hasan approximately similar objecti,'e, except rhat it is the '\\al' of seeing' of acultural system, rather than a historical period, which has to be elucidated.

I ha\'c no objection to Price's suggestions so far as increasing the recogni­tion afforded to non-Western art and artists is concerned. Indeed whal lI'ell­intentioned person could object to such a programme, except possibly the'connoisseurs', who derive a reactionary satisfaction Irom imagining Lhal theproducers of the 'primilive art' they like to collect are primeval savages, barelydescended from Ihe trecs. Thcse idiol'S c.1n be dismissed out of hand.

All the same, I do nOl lhink that the c1ucidal'ion of "Qll-I1'~SU'" a~sfh(1;(s)'s­lI!1I/S consti£uICS an 'anthropology' of art. Firstly, such a programme is cxclus­i,·c1y cultural, rather th:an soci:al. Anthropology, from my poim of view, is a

social science discipline, not a humanity. The distinction is, I admit, c1usi\·e,but it does impl~ rhar the 'anrhropology of art' focuses on the social comexrofan production, circulation, and receplion, rather than the e\'aluation of par­ticular works of art which to my mind, is the function of a critic. It may beinterestmg to know why, for example, rhe Yoruba evaluate o~'ing as aes­thetically superior to anorhef (R. F. Thompson 1973), but that docs not tell usmuch about why rhe Yoruba carve 10 begin with. The presence of large num­bers of can'ings, can'crs, and critics of carvings in Yorubaland al a certainperiod in time is a social faL1: whose explanarion does nllt lie in the domain ofindigenous aesthetics. Similarly, our aestheric preferenct:s cannot by them­selves account for the existence of the objects which we assemble in museumsand regard aeslhetically. Aesthctic judgements are only interior mental acts;art objects, on the othcr hand, afe produced and circulated in the ext'ernalphysical and social world. This produclion and circulatiun has to be sustainedby certain social process(.'S of an Ilbjecri\"C kind, which arc connected to olhersocial processes (exchange, politics, religion, kinship, etc.). l;nlc~--s, for instance,there were SC1:rct societies such as Poro and Sande in West Africa, there wouldbe no Pom and Sande m.:l.Sks. Poro and Sande masks can be regarded and eval­uated aesthetically, by ourseh·es, or by the indigenous an public, only bcL'3USCof the presence of certain social institUTions in that region. E\'en if one were toconcede thaI something akin to 'aesthetics' exists ;\s a feature of the ideationalsystcm ofevery culture, one would be far from possessing a theory which couldaccount for the production :and circulation of particular works of art in par­ricular social milieux. In f3Ct, as I ha\e argued elsewhere (Gell 1(95), I am farfrom convinced that e\:ery 'culture' has a component of its ideational systemwhich is comparable to our own 'aesthetics'. I think that the desire to see theart of other cultures aesthetically tells us more ,lbout our own ideoIOb,)' andits quasi-religious vencration of art objects as aesthetic talismans, than it docsabout these other cultures. The project of 'indigenous aesrhetic~' is essent'iallygeared to refining and expanding the aesthetic sensitivities of the Western artpublic by providing a cultural context within which non-Western art objectscan be assimilated to the categories of Westl:rn aesthetic art-appreciation. Thisis not a bad thing in itself, but it still falls far short of being an anthropologicaltheory of art production and circularion.

I say this for reasons that are unafTecled by the L"OrrL'CtnCSS or orherwise ofm~ views about thc impossibility of using 'aesthetics' as a universal p:lfametcrof cultural description and comparison. Even if, as Price, Coote, lVlorphy, andolhers suppose, all cullures have an \Icsthetic', descriptive accounrs of other~ultures' aeslhetics would not add up to an anthropologicalthcory. Distinct­l\ely 'amhropological' ("heories have certain dcfining characteristics, whichthc.."C accounts ofC\-a!uati\·c schemes \\ ould lack. Evaluati,'c schemes, of what­e\"(:r kind, are only of anthropological interest in SO far as they playa part withinsocial processes of interaction, through which Ihe~ are generated and sUSl':lined.

J .2. The An Objecl

providing a 'L'Onten' for non-Western art such that this an can become aeec.....sible to a Wc...tcrn art-public. Ilowe\'er, the responses of the indigenous art'public' to indigenous art is hardly exhausted by the enumeration of those con­texts in \\hich something like an evaluative acsthetic scheme is deployed in'appteciating' art. Such contexts may he rare or non-existent, yet 'whaT looks10 us like an' is none the less produced and circuhnerl.

A purdy cullural, aesthetic, 'appreciative' approach to an objccts is ananthropological dead end. Instead, the question which interest<; me is the pos­sibility of formulating a 'theor~ of an' which fits naturally into the context ofanthropolog:r, ghen the premiss that anthropologiL-a1 theories are 'recogniz­able' initially, as theoriL'S aboUl social relationships, and not anything else. Thesimplest way to imagine this is to suppose thai there could be a species ofanthrupologic'll theory in WhiC~~f{.~sociaLagffiti ~e,~ certain 'O~­texts, substituted l~r-lrfi1r/ijb;e[t!i:

-=- ----==::;::

This immediately raises the question of the definition of the 'art objecl', andindeed, of ,an' itself. Howard '\'lorph) (II)<).I.: 648-85) in a recent discussion ofthe problem of the 'definition of art' in the anthropological context, considers,and rejects, the (Western) institutional definition of art, that. 'art' is whate\'eris treated as an by members of the institutionally rel."ognized an world (Dan to

11)64)-critics, dealers, collectors, theoreticians, crc. This is fair enough: thcreis no 'an world' to s[K:ak of in many of the societies which anthropologistsconcern themselves with, yct these societies produce works some of which arerecognized as 'art' by our 'art world'. ALx:ording to the 'institutional theory ofart', most indigenous art is only 'art' (in the sense we mean by 'art') beclusewe think it is, not because the people who make it think so. Accepting the aTlworld's definition of an obliges the anthropologist to bring to bear on rhe artof other cultures a frame of reference of an ovcrtly metropolit:lI1 character. Tosome .e~t~nt thi~..jnevitab~aIllhropology is ~ l~ct~opolitan activity, juslli.keart CtlllClsm) bitt J\~orphy I~Jndersl'an~ablr dlSlllchned to aLx.-cpt the verdIctof the (amhropo1bgicaUy-umnformcd) Western art "orld as to the definition of'art' beyond the physica.l frontiers of the West. He proposes, instead, a dualis­tic dcfinition; art objecls arc those 'ha\'ing semamic and/or aesthetic properties~hat arc uscd lOr presentatloiial or represemational purposes' (iBid. 655), thatI:', Cit f~er art objCCtS:irCsign-vehides, Con\'ey-ing 'rncanin-g'. or the~~are_obje(,;tsmade 111 order to provoke :l culturally endorsed aestheTic response, or both ofthese simuhancoiISly":- - _

I find-both of thc...c conditions for an object status questionable. I havealready expresscd the opinion thaI 'aesthetic properties' cannor be abslTa(,;redanthropologicall~, from Ihe social prOCt:SSCS surrounding the deplovmellt ofcandidale 'an objt:cts' in specific social settings. I doubt, for examp·le. thaI a

5Thr Problrm Drjillrd

I

-I nit Proh!'!n DrjirttJ

The anthropology ofla\\, for instanl-c, is not the study of legal-ethical principles-other peoples' ideas of right and wrnng-but of disputes and their resolu­tion, in the course of which disputant... do often appeal lO such principles.Similarly, the anthropology of art c:lOnot be the study of tbe ac... thetic prin­ciples of this or that culture, but of the mobilization of aesthetic principles (orsomething like them) in the wurse of social inTeraction. The aesthetic theoryof an just does not resemble, in any salient respe<:t, any existing anthropolo­gical theory aboUT S()(.;al pnK:esses. What it resembles is existing Western artth(:ory-"hich ofcourse it is, applied no longer lO 'Western' an, but to exoticor popular arl. To de\'e1op a distincti\'e1y anthropological thcor~ of art it isinsufficient to 'borrow' existing an theory and apply it lO a new object; onemusl develop a new variant of (xisting o1/thfflpological th(ory, and ~lpply it TO artIt is not that I WflnT to be more original than my colleagues who have appliedthe existing theory of art to exotic objects, I just \\,;1111 TO be unoriginal in :1

new way. 'Existing anthropological theories' are nOI about art; they are aboutwpies like kinship, subsistence economics, gendlT, religion, and the like. TheobjL'Cti\-e, therefore, is to neate a theory about art which is anthropologicalbeo1usc it resembles these other theoriL'S that one cm confidcntl~'describe asanthropological. Of course, this imitativc strdtegy vcry much depends on whatson of a subject one considers anthropology to be; and how this subject differ.;from ncighbouring ones.

-WhaT constituteS the defining char,H,:teristic of 'anthropological theories',as a class, and what grounds havc T for asscTling that codifying acsthetic­evaluative sehemes would nor fall under such a ruhric? My view is that in sofar as anthropology has a specific subject-maner at all, that subjL'1.:1-mancr is'social relationships'-rclationships between particioants in social systems of\1irious kinds. I recognizc that many anthropologists in the tradition of Boasand Kroeber, Pri,e among them, consider that the subject-maller ofanthropo­logy is 'culture'. The problem with this fotmulation is that one only discoverswhat anybody's 'culture' consists of by obsen'ing and recording their culturalbehaviour in some specific sclting, that is, how the)' relate to specific 'others'in social interactions. Cuhure has no existence independently of irs mani­cstations in social interactions; this is true even if one sirs someone down andsks them to 'tcll us about \our culturc'-in this case the interaction in ques­ion is the one berwL'Cn thc· inquiring anthropologist and the (probably rather

pemused) informant.The problem with the 'indigenous aesthetics' programmc, in my view, is

that it tends to reify the 'aesthetic response' independently of the social confextof its manifestations (and that Boasian :lnthropology in general reifies Clliturc).In so far as Ihere can he an anthropological theory of 'aesthetics', such aIhL'Ory would tr) to explain why social agents, in particular scttings, producethe responses lhat they do to particular worl.:s of art. I think that this canhe distinguished from the laudable, but essentially non-anthropological task of

warrior on :\ banlefidd is 'aesthetically' interested in the design 011 an opposingwarrior's shield' ret il was so as to be seen by this warrior (and £0 frighten him)that the design \~';''i placed (here. The shield, ifil resembles lhe one in the fromis­piece (p. xxiv), is indisputably :\ work of 311 of the kind inc.cresting to t~e

anthropologist, but its aesthetic properties (for liS) 3fC toTally IrI'c1~\'an[ t~) Itl'i

anthropological implications. inthropologlcally, it is not a 'bcauuful' shield,

\'\UI a fear-inducing shield. The innumerable shades ofsociaVemotional n:sponscsto artefacls (of terror, desire. awe, fascio;ltion, etc.) in-the unfolding patternsof social life cannot be encompassed or rcdUl.:cd (0 aesthetic feelings; not with­

jout making the aesthetic response so gencralized as to be al.tog.cthcr meaning-less. The effect of the 'aestheticization' of response-theory IS SImply to equatethe reactions of the ethnographic Other, as far as possible, to our own. In fact,responses to artefacts are ne"er such as to single out, among the spectrum ofavailable anefacls, those thal are auended to 'aQithetically' and those that are

noloNor am I happy with the idea that the work of aJt is recognizable, generic­

ally, in that it participates in a 'visual' code for the.E?m~unication,ofme~mi;?I entirely reject lhe idea illatanytfiing, except language It'ielf, has meamng In

the inte~ded sense. Language is a unique illsbtution (with a biological basis).Using language, we can talk about objects ami attribute 'meanings' to them inthc sense of 'find something to say about them' but visual art objects arc notpart oflanguage for chis reason, nor do [hey conslitute an altcrnative language.Visual art ohjcCls are objt.'Cts about which we may, and commonly do, speak­but the\' themselves eilher do not speak, or tbey uttcr natural language i.ngraphe~lic code. We talk about objects, using signs, but art objccls arc nor,except in spech1l cases, signs themselves, with 'meanings'; and if they do havemeanings, then they arc parI O//llllguagt' (i.e. graphic signs), not a separatc'\'isual' language. I shall rerum to this subject at interyals, since my polemica!r.linst the idea ofa 'language ofan' has many diITerent aspects to it, which areb~ner dealt with separately. For thc present,let me simply warn the reader th~tI ha,'e a\oided the usc of lhe notion of 'symbolic meaning' throughout lhlswork. This refusal to discuss :Irt in tenns of sym s and-meanings may occa­sion some surprise, since the domain of'art' and the symbolic arc held by manyto be more or less coextensive. [n place of symbolic communication, I pillce allthe emphasis on agmcy, i,,/rtl/ioll, (ausa/ioll, r~su/f, and ITalls/orlllalioll. I vicwart as a systelli"'Of:\ct~dea to change lhe world rather than ~n~ode sym­bolic propositiuns about it. The 'action'-centred approach to arllS m,h~rcnt'y

\..... more anthropological than the alrernati,'c scmiotic approach bccau~e It IS pre-, ,f occupied with the pracliL'a1 mediatory rolc of art objects in the SOCIal process,

J,fO-JJ~ T'".Ither than \\ith the interpretation of ohjects 'as. if' .th:y ".ere tcxts. ,l Having rejected lVlorphy's twO criteria for dlscnmlllatmg the class ~f art

I I - "I oscs of the alllhropologv of art. I am. of course, stilI lefto ljcctS lor I le purp . • '.. C b-with the unsoh'cd problem of proposmg a entenon lor art 0 Ject Slatus.

6 The Problem Jkjind rhe Proh/em LkjineJ i

rortunately, howe\'er, the anthropological thcory of art doc... not need to pro-­\·ide a nitcrion for art object status which is independent of the theory itself.Th.c anthropologist is. n.ot obliged to define the art object, in adnnce, in a waysatlsfacwry to aesthetlclans, or philosophers, or .11"t historians, or anybody t:Ise.The definition of the an objcct I make use of is not institutional, nor is it ac...­thetic or semiotic; lhe definition is Ih~orefi(QI, The art object is whate\'er isinserted into lhe 'slot' prm'ided for art ohjects in the syslem of terms and rela­tions envis'lged in [he theory (10 be outlined later). l\·othing is decidable inadvance about Ihe nature of this objl.'Ct, bee.lUse the theory is premised 011 theidea that the nature of the art object is a function of the social-relational matrixin which it is embedded. II _has no 'intrinsic' nature, independent of the rela­tio",11 conte~. Most of the art objects I shall ;ctually discuss arc wcll-knowno~ we have no difficulry in identifying as 'art'; for instance, the IVlona \t\~isa .. lnasmuch ~ we recogniz~ a pre-theoretical category of art objcct's~split .jI1lto the t\\'o maJor subcawgoTlcs of 'Western' art objects and 'Indigenous' or , f'Ethnographic' art objects--:-l conduct the ~Iiscussion in terms of'prototypk'al' <~members of these CiUegoTlQi, for convenience's sake. But' in faa anvthin.... ;I'

whalsoever could, I."flllceivably, be an art object from the anthropologicai~of v.iewL.!ncluding living persons, because the anthropological theory of art(whl~h.we can, roughly define as lhe 'social rclations in thc \·icinity of ubjectsm~Ja(mg SOCIal agency') mer:gcs seamlessly with the social amhropologv ofpersons. an<l.their bodiQi. Thus, from the poim of vicw of the anthropology ofart, ~n Idol III a temple belicved to be the body of the divinity, and a spirit-mcdlum, who likewise provides the divinity with a lemporary bod" are treatedas thcoretk'ally on a par, despite the fact th:n the former is;o art~fact and thelatter is a human being.

1·3· Art Sociology '" \d'

[ have jusl pro\'isionally defined the 'an~hropoloi of art' as the theoretical~tUd~' of 'social relatilln~ in the vicinity of objccts l~ediating social agency' and

ha\e sugg~cd.that III order for the amhropology of art til be specificallvanthropolopcal" It has to proceed on the basis that, in rele\'ant thl.'OreliC;;1respects, art objects arc the cqui\'alenl of pcr:o>ons, or more precisely socialagents: Arc there no alternativcs to this seemingly radical propositio~? Well,o~e might draw bad from the abyss, and agree that evcn if thc anthropolo­gl~1 theor) .ofart were not to be 'cross-cultural acsthetics' or a bmnch ofsemi­Otics, th~n ~t could still be a sociology of art 'institution.~' which would nOIneccss~nl~' In\·oh·c the radical aflirmation of the personhood of art ob'eetsT~ere Is,.md:ed; a flourishing 'sociology ofan' which concerns itself pre~isel\;\\ tth the mstnut!onal pammeters of art production reeeprion and circulatio~H~wc\'.c~, it is not coinc~dentallhal the 'sociology o'f art' (insti; utions) has bt:e~prtmartl~, concerned wuh Weslern art, or, failing [hat, the art of advanced

8 Th~ PTOb/~1I/ Drfilld Tht Probltm Ikjim:d 9

states with bureaucracies, such as China, Japan, Ctc. There r..'llnnot be an 'insti­tutional' sociology of art unless the relc\':mt institutions arc extant; thal is, anaft public, public or private p:ltronagc of artists, art critics, ;ll'[ museums,academies, art schools, and so on.

Writers who dl"aj with the sociology of art, such as Berger (1972) andBourdicu (1968, 1(}84), concern rhcmscln:s with particular institutional char­acteristics of mass societies, r.llher than with the network of relationshipssurrounding particular artworks in specific interactive scnings. This divisionof labour is characteristic; anthropology is more concerned with the imme­diate context of social interactions anti their 'personal' dimensions, whereassociology is more preoccupil--cl with insritutions. Of course, there is continuitybetwl-'Cn the sociological/institutional perspective and the anthropological/rela­tional one. f\nthropolngists cannot ignore institutions; the anthropology of arthas to consider the institutional framework nf rhe produl.:tion and cin:u!atinnof :lrrworks, in so far as such institutions exisr. But it remains true to say Lhatthere arc many societies in which the 'institutions' which provide the contextfor the production and circulation ofan are not spcl-;alized 'art' institutions assuch, but institutions of more general SOlpe; for example, CUlL", exchange sys­tems, etc. The anthropology of :lrt would forever remain a \"ery undevelopedfield were it T.O restrict itself to institutionalized an production and circul:uioncomparable to that which can readily be studied in the context of advancedbureaucraticfindustrial states.

As it is, the 'sociology ofart' is represented in the 'anthropology ofart' prim­arily in the guise of studies of the market in 'ethnographic' art, such as thedistinguished recent work by Steiner (1<)<J4)..1\'lorphy (191)1), Price (1989),Thomas (I991), and others have wrinen very illuminatingly about the recep­tion of non-Wesfern art by the Western art-public; bUI these studies concernthemseh'cs with the (institutionalized) art world of the West, and the responsesby indigenous people to the reception of their artistic pmduction by thisalien an world. I think one can distinguish bct\\een these invcstigations of thereception and appropriation of non-Western art, and the scope of a gcnuinelyanthropological theory ofart, which is 110t to denigrate such studies in any way.One has to ask whether a given work of art was actually produced with thisreception or appropriation in mind. In the contemporary world, much 'ethno­graphic' art is actually produced for the metropolitan market; in which casethere is no possible way of dealing with it except in this specific frame\\'or~.

However, it also rem,lim true that in the past, and still today, art was and IS

produced for much more limited circulation, independently of :ll1Y receptionwhich mar be accorded to this art across cultural and institutional boundaries.These I~I COll[exts, in which art is produced nOI as a function of the existenceof specific 'art' inslitutions, bOl as a b)-produCI ofrhe mediarion of social lifeand the existence of insritutions of a more general-purpose kind, justify the

assertion of at least rclati,'C autonomy for an anthropology of art nor circum­scribed by the prescnce of institutions of any specific, art-related, kind.

It seems, then, that the anthropology of art can be aT least provisionally sep­,lratcd from rhe study of art institutions or the 'art world'. \Vhieh implies theneed to return to, and ret:onsider, the proposition ad\'anced abore. To suggestthai art objl-'ClS, to figure in an 'anthropologicr1' theory of art, ha\"e to be con­sidered as 'persons', seems a bizarre notion. But only if one fails to bear in mindthat the emire hisrorkal tendency of anthropology has been towards a radicaldcfamiliarization and relativiz;ttion of the notion of 'pcrsons'. Since the outsetof the discipline, anthropology has been signally preoccupied with a series ofproblems to do with oSlensibly pl-'Culiar relations berwl-'Cn persons and 'things'which somehow 'appear a.,,', or do dury as, persons. This basic theme \\35 ini­tially announced by Trior in Primiliu CII/tllrt (1875), where, it will be recalled,he discusses 'animism' (i.e. the attribution of life and scnsibilitv to inanimatethings, planl~, animals, etc.) as the defining attribute of 'primitive' culture, ifnot culture in general. frazer rcrurns to precisely the same theme in his volu­minous studies of sympathetic and contagioU'; magic. Identical prl-'1>ccupationssurface, in a different way. in Ihe work of ~1alinowski and Mauss, this time inrelation 10 'exchange' as well as the classical anthropological theme of magic,abOllt which each also wrote eXlensi\'ely.

The proposition just advanced, that the anthropological theory of art is thetheory of art which 'considers an obj(.'Cts as l>ersons' is, I hope, immediatelyand legibly ,Maussian. GiYCn that prestations or 'gifts' are treated in Maussianexchange theory as (extensions of) persons, lhen there is ob\iously scope forseeing art objects as 'persons' in the same way. In fact, il might not be goingroo far to suggest that in so Elr as Mauss's theory of exchange is the exemplary,pro.totypical, 'anthropulogicaltheory' then the way to produce an 'anrhropo....10gK":lltheory of art' would be to construct a theory which rescmbles :\1auss's,but which \\as about art objects rather than prestatiolls. Levi-Srrauss's kinshipIheur~' is Mauss with 'prestations' replaced by 'women'; the proposed 'anthro­po~oglcal theory of an' would be iVlauss with 'prestations' replaced by 'artobJccts'. Actually, this would bc a tra\'esty of the theory I am about to produce,but I make the analogy in order to guide the reader as to mv basic intentions.The point I am making is rhat an amhropological theory of'any giwn topic isonl~ 'amhr~pologi'-'ll' to the extent lhal it resembles, in key respects, ot/faamhr?pologlcal theories, olherwise lhe designation 'anthropological' has nomeal.H.ng. M) aim is lo produce an anthropological theor)' of art which hasafflll111e~ towards other amhropo!ogic;ll theories, not just l\huss's of course,but vanous others as well. One of my basic objections to the '(;l'os.."-culturala~l~~tic.,,' and 'semiotics' rhcories of ethnographic art is that the theoreticalaffill1tles of thl-'S: a~proaehes lie in (Western) aesthetics and art-theory, noraUlonomously wlthm anthropology itself. II may he Ihal there is no useful

The Them)' IIIthe Arl Ne.r". 13

2

The Theory of the Art Nexus

2. I. Constructing a Theory: Terms and Relations

To construct such a theOT\' it is first of all necessary to define certain theoret­ical emities (tenus) and r~lalions. Just now, I suggeslcd thaI sUI.:h a theorywould 'look like' familiar amhropolO',!:ical theories, such as the theory ofexchange, or the theory of kinship, but that it would replace some of the termsof such theories with 'art objects',

J-Iowner, this raises immediate diffIculties, in that. 'art objects', 'works ofart", Of 'artworks' may form a readily identifiable class of objects in some artsystems, but this is hardly true of ,Ill of them, especially not in ant~ropologicalcontexts. In effect, if we make 'the work of art' the corner-stone of the anthro­pological theory of art, the theory itself becomes instantly otiose, for reasonswhich ha,'e alrc..'tdy been aJluded 10. To discuss 'works ofarl' is 10 J.iscuss enti­ties which have been given a prior iltstitutio/lal definilion as such. The institu­tional recognition (or 'enfranchisemcnt') of art objects is the subject-maner ofthe sociology of art, which deals with issues which arc complementary to theanthropology ofan, but do not coineide with it. Ofeoursc, some (in facl, many,or even aU) of the objects which fall within the scope of the sociology ofart mayalso be considered 'anthropologi,,:ally' as entities in whose neighbourhoodssocial relationships are formed; but 'work ofart' status is irrelevant to this. Theanthropology of art, if it is to be distinguished from the sociology of art, can­not restrict its S(:upe ro 'official' art institutions and recognized works of art. Itcannot, in fact, talk about 'works of art' at all, not only because of the institu­tional implications of 'work of art' status, but because this term has undcsir­abl\' exclusive connotations. An object which has lx.'Cn 'cnfrnnchised' as an artobj~ct., becomes an art objcct exclusively, from the standpoint o~th~'Ory,and ~-anonl\' be discussed in t.erms of tbe parameters ofart-theory, which IS what bemg'enfranchised' in this way is all.thOuI. The anthropological theory of arl can­not aITord to have as its primary rheoretical term a c;ltegory 01' taxon?f objectswhich are 'exclusive!v' art objects because the whole tendency of this thea?,as I have been sugges~ing, is to explore a domain in which 'objects' merge With'pcople' by virtue of the exislcn~c of: social relalions hetween persons andthings, and persons and persons VIII th111gs. ..

I do not promise never III menlion art objects ag-Jlll; 10 facl, J sh~l1 d~ s.o"1><::ltl..-(1)' sine!: excessive terminological consistency is the enemy of I~telhgl-

, . 1 • b 'bilil), my primary objective here. Bur I do not mtem 10 use arlO ,eet or

'\\·ork of art' or 'artwork' as technical terms, nor to discuss when an object isan 'art object' and when il is somelhing else. The technicallerm I am going to

employ is 'index'. This requires explanation.

2.2. The Index

The anthropology of art would not be the anthropology of urI, unless it wereconfined 10 the subset of soci21 relations in which some 'object' were rdated toa social agent in a distincl'h'e, '2rt-like' way. We ha\'e dismissed the idea tharobjects arc related to social agents 'in an an-like way' if (and only if) socialagents regard these objects 'aesthetically'. But in lhis case, what alternativemeans can be proposed to distinguish art-like relations bet\\een persons andfhings from relarions which arc not art-like? To simplify the problem, I shallhenceforth confine the discussion to the instance of \'isual art, or at least, 'vis­ible' art, excluding verbal and musiC'J.I an, though I reeogni'l.e that in practicethese arc usually inseparable. So the 'things' of which I speak may be under­stood to be real, physical things, unique and identifiable, not performances,readings, reproducrions, cTt. These stipulations would be out of place in mostdiscussions of art, but they are necessary here if ollly because difficulties canbest be surmounted one at a timc. And it cenainly is very difficult to proposea criterion which would disringuish the typcs of social relations falling underrhe scope of the 'anthropology ofart' from any orhcr social,relations.

I propose thar 'an-like situations' can be discriminated as those in which thematerial 'index' (the visible, physical, 'thing') permits a parricular l:ognitiveoperation whieh I identify as till' llMue/ioll ofagmey. An 'index' in Pien:cansemiotics is a 'natural sign', that is, an entity from which the ohscrver can makea rausal i'1frr~nu of some kind, or an inference about the intentions or (:ap­abilities of another person. Thc usual example of an 'index' is visihle smoke,betokening 'fire'. Fire C'"Juses smoke, hence smoke is an 'index' of fire. Anothervery common example of an index is the human smile, indexing a friendly ani­tude. Howe,·cr, as we all know, smokc can arise in the absence offirc, and smilesmay decei\'e. The oognitive operation through which we infer the presence oftire (given smoke) or friendliness (givcn lhe smile) is not like the cognitiveopemtiun by means of which we 'Imow' thar 1; + 1; = 4, or that if somebodyutters the word 'dog' he means 'C'J.nine' and not 'railway train' or 'bunerfly'.Indexes are not part of a c;tleulus (a set of tautologies, like mathematics) nor are~hey components of a mllural or artificial language in which terms have mcan­Ings establishcd by convention. Nor arc inferences from indexes arrived at byinduction or deduction. We have ,not made a test, and established that by :1 Ia\~o.f nature, smoke means fire. In tact, we know that smoke may not mean fire,Since \\·c know of fire-less ways of producing smoke, or rhe appearance ofsmoke. Since smoke as an index nfllrc docs nOI follow from anv known law ofnature, dcduCI'i\'c1y or inductively arrived at, and ;s neither a ;;lUtology nor a

theory of art that cm be founded on or derived from existing anthropologicaltheory, but umil one has made the cxpcrimem to construct a genuinely anthro­pological theory of art, this question cannot be decided.

I -4- The SillUJlletfe oIo" Alllhrop%gical Theory

The position I htl\'C reached is that an anthropological theory of art is onewhich 'looks like' tm anthropological theory, in which certain of the ,daM,whose relations aTC described in the theory, arc works of arl. But .....h31 do'anthropological' theor;~ look like? Can ooe really gi\'c an idcntificalOry sil­houclIe of an anlhropologi<:'31 theory as opposed to any other kind of theory?Well, possibly nOI, in that anthropology is a brood church and is only \"cryambiguously distinct from other disciplines, such as sociology, hislOry, socialgeography, social and cognitiyc psychology, and so on. This much I readilyconcede. On the other hand, let us considerl\,hat anthropologist.'i do bc.'it from

rE;C viewpoint of neighbouring disciplines. nthropolog)' is, to put it bluntly,nsidcrcd good at providing close-graine . analyses of Qfp"a!!!!!!y irrational

beha\'iour, performances, utterances, etc. (The 'my brother IS a green parrot'problem: Sperber T985; Hollis 1970.) Since almost all behaviour is, fromsomebody's point of view, 'apparently irrational' anthropolog)' h;lS, pos.'iibly,a secure future. How do anthropologists solve problems about the apparentirrationality of human beha\'iour: They do so by locating, or contcxtualizing

\ beha\'iour not so much in 'culture' (which is an abstraction) as in the dynam-

)

icS of social interaction, which may indeed be conditioned by 'culture' hUIwhich is better seen as a real process, or dialectic, unfoldin~~imc...J'hcamhropologi~'al intc:pretative PC~peCTrve on soci31. beh~1\'iour. is shar~d, ~~~d­less 10 say, wnh SOCIO!Ob'} and SOCIal psychology, nol TO mentIon othcr diSCIp­lines. Anthropology diITers from these in providing a particular depth oIfi)(/I$,which pcrh,lps onc could hest describe as 'hiographical', thai is, the \'iew taken

Jby anthropology of social agenls attempts 10 replicate the time pcrs\>ecTive ofjthese ag,ents on .themselves,. whereas (h.i~torical) sociology is.often, ~) 10 s~~k,sllpr~bJ.O:~Ki1.' and SOCIal or cognmvc psychology are mfra-blographlcal.

~ " Anthropology therefore tends to focus on the 'act' in the context of rhe 'life'­...... l_or morc precisely, th~ 'stagc ofJife'---of the agenl. The fundamcntalycriodic­

iryof-anthropology is the life cycle. This time perspective (fidelity to ,thc~ biogr;lphieal) dicrates just how close to and .how fa~ away from ~I~c subject~ th~ anthropol()~is\' STands; if the anthropolo~l.st st~~~es (sa}? co~m.tl~l~ al .t~c

). Illlcro-sealt:: Iyplcal of mm:h laboratory cogl1ttlve ps)'cholog}.' the blO~~,\phlcal.,; i perspective is lost ,\Ild lhe anlhropologist, in effect, i~ just dOing cogllltlve psy-\... cholog}'; conversely, if the anthropologist's perspccfl.ve.cxpands ro the degr.ee~ that the biographic,ll 'life cycle' rhythm no longcr deJlln1ts the scope ofthc dts-1 course, he or she is doing history or sociology .

.~ ~j "~) ?

Thl ' Probkm Defined 11

Perhaps this definition of anthropology will not be to the liking of all, but Iwould argue that il docs in fact encumpass most of the work which is regardedas typically 'anthroJlological'. 'I 'his specifically biographical depth of focus hasofcourse also a spatial correlate; the Sp,lCes ofanthropology are those which aretraversed by agents in the course of their hiographies, be they narrow, or, as isbecoming increasingly lhe case, wide or evcn world-\\·ide, Moreo\'er, it dictatesa {.-ertain \'iew of social rel:uiOIlS. Anthropologists typically view relationshipsin a 'biographical' context, by which I mean that relationships are seen as partof a biographical series entered into :It different phases of the life C\'dc'Sociologil..'al' relations, are, so to speak, perennial, or supra-biographiC:l.l: lik~the relalion between The classes in c:lpitalism, or the relationship betweenstatus-groups (castes) in hierarchical societies. 'PsychologiC:l.I' relations, on theother hand, arc infra-biographical, often no more than momentary 'encoun­ters', as, for instance, in experimental settings in which subjects arc requiredto interact with one another, and with the eXlx:riml--ntcr, in ways which ha\'eno biographiC:l.1 precedems or consequences, i\nthropologiC:l.1 relationshipsare real and biographically consequential ones, which articulate to the agent'sbiographical 'Iifc prnject'.

If these stipulations arc correct, then the chardcteristic silhouette of an'anrhropologiC:l.l theory' is beginning to emerge. Anlhropological theories aredistincti\e in that they are t)'pic::all) about social relationships; these, in turn,occupy a certain biographk'al space, O\'cr which culturc is picked up, {rdnS­fanned, and passed on, through a series of life-stages. "'he swd\' of relation­ships oyer the life l..'Ourse (the relationships through which cultu~e is acquiredand reproduced) and the life-projects which agents seek to re-dlize throughtheir relations with others, ;tllo\\'s anthropologists to perfonn their allottedintellectual Task, which is to explain why IX:oplc behavc as thcy do, even if thisbehaviour seems irralionaL or cruel, tIT amazingly saintly and disinterested,~s may be, The aim of anlhropological thenry is to make sense of behaviourIn the context of social relations. Currespondingly, the objccli\'C of the anthro­pological theory of art is to account for the produClion and circulation of anobjects as a function of this relational eontcxt.

,-- bI~C'{J(

.-:) __ t,?-L

II ...) cf.-( tJ-. J

cce1

The Problem Dejii/ed10

The Them)' IIIthe Arl Ne.r". 13

2

The Theory of the Art Nexus

2. I. Constructing a Theory: Terms and Relations

To construct such a theOT\' it is first of all necessary to define certain theoret­ical emities (tenus) and r~lalions. Just now, I suggeslcd thaI sUI.:h a theorywould 'look like' familiar amhropolO',!:ical theories, such as the theory ofexchange, or the theory of kinship, but that it would replace some of the termsof such theories with 'art objects',

J-{owncr, this raises immediate diffIculties, in that. 'art objects', 'works ofart", Of 'artworks' may form a readily identifiable class of objects in some artsystems, but this is hardly true of ,Ill of them, especially not in ant~ropologicalcontexts. In effect, if we make 'the work of art' the corner-stone of the anthro­pological theory of art, the theory itself becomes instantly otiose, for reasonswhich ha,'e alrc..'tdy been aJluded 10. To discuss 'works ofarl' is 10 d.iscuss enti­ties which have been given a prior iltstitlltio/lal definilion as such. The institu­tional recognition (or 'enfranchisemcnt') of art objects is the subject-maner ofthe sociology of art, which deals with issues which arc complementary to theanthropology ofan, but do not coineide with it. OfeouTSC, some (in facl, many,or even aU) of the objects which fall within the scope of the sociology ofart mayalso be considered 'anthropologi,,:ally' as entities in whose neighbourhoodssocial relationships are formed; but 'work ofart' status is irrelevant to this. Theanthropology of art, if it is to be distinguished from the sociology of art, can­not restrict its S(:upe ro 'official' art institutions and recognized works of art. Itcannot, in fact, talk about 'works of art' at all, not only because of the institu­tional implications of 'work of art' status, but because this term has undcsir­abl\' exclusive connotations. An object which has lx.'Cn 'cnfrnnchised' as an artobj~ct., becomes an art objcct exclusively, from the standpoint o~th~'Ory,and ~-anonl\' be discussed in t.erms of (be parameters ofart-theory, which IS what bemg'enfranchised' in this wav is all.thOuI. The anthropological theory of arl can­not aITord to have as its primary rheoretical term a c;ltegory 01' taxon?f objectswhich are 'exclusive!v' art objects because the whole tendency of this thea?,as I have been sugges~ing, is to explore a domain in which 'objects' merge With'pcople' by virtue of the exislcn~c of: social relalions hetween persons andthings, and persons and persons VIII th111gs. ..

I do not promise never III menlion art objects ag-Jlll; 10 facl, J sh~l1 d~ s.o"1><::ltl..-(1)' sine!: excessive terminological consistency is the enemy of I~telhgl-

, . 1 • b 'bilil), my primary objective here. Bur I do not mtem 10 use arlO ,eet or

'\\·ork of art' or 'artwork' as technical terms, nor to discuss when an object isan 'art object' and when il is somelhing else. The technicallerm I am going to

employ is 'index'. This requires explanation.

2.2. The Index

The anthropology of art would not be the anthropology of urI, unless it wereconfined 10 the subset of soci21 relations in which some 'object' were rdated toa social agent in a distincl'h'e, '2rt-like' way. We ha\'e dismissed the idea tharobjects arc related to social agents 'in an an-like way' if (and only if) socialagents regard these objects 'aesthetically'. But in lhis case, what alternativemeans can be proposed to distinguish art-like relations bet\\een persons andfhings from relarions which arc not art-like? To simplify the problem, I shallhenceforth confine the discussion to the instance of \'isual art, or at least, 'vis­ible' art, excluding verbal and musiC'J.I an, though I reeogni'l.e that in practicethese arc usually inseparable. So the 'things' of which I speak may be under­stood to be real, physical things, unique and identifiable, not performances,readings, reproducrions, cTt. These stipulations would be out of place in mostdiscussions of art, but they are necessary here if ollly because difficulties canbest be surmounted one at a timc. And it cenainly is very difficult to proposea criterion which would disringuish the typcs of social relations falling underrhe scope of the 'anthropology ofart' from any orhcr social,relations.

I propose thar 'an-like situations' can be discriminated as those in which thematerial 'index' (the visible, physical, 'thing') permits a parricular l:ognitiveoperation whieh I identify as till' llMue/ioll ofagmey. An 'index' in Pien:cansemiotics is a 'natural sign', that is, an entity from which the ohscrver can makea rausal i'1frr~nu of some kind, or an inference about the intentions or (:ap­abilities of another person. Thc usual example of an 'index' is visihle smoke,betokening 'fire'. Fire C'"Juses smoke, hence smoke is an 'index' of fire. Anothervery common example of an index is the human smile, indexing a friendly ani­tude. Howe,·cr, as we all know, smokc can arise in the absence offirc, and smilesmay decei\'e. The oognitive operation through which we infer the presence oftire (given smoke) or friendliness (givcn lhe smile) is not like the cognitiveopemtiun by means of which we 'Imow' thar 1; + 1; = 4, or that if somebodyutters the word 'dog' he means 'C'J.nine' and not 'railway train' or 'bunerfly'.Indexes are not part of a c;tleulus (a set of tautologies, like mathematics) nor are~hey components of a mllural or artificial language in which terms have mcan­Ings establishcd by convention. Nor arc inferences from indexes arrived at byinduction or deduction. We have ,not made a test, and established that by :1 Ia\~o.f nature, smoke means fire. In tact, we know that smoke may not mean fire,Since \\·c know of fire-less ways of producing smoke, or rhe appearance ofsmoke. Since smoke as an index nfllrc docs nOI follow from anv known law ofnature, dcduCI'i\'c1y or inductively arrived at, and ;s neither a ;;lUtology nor a

Thr 11,rory Qfthr Art Ntxus

convention oflanguage. we need another technic-J.I term to designate rhe modeof inference (or cognitive operation) we bring to belr on indexes.

2.3. Abduction

The term employed in logic and semiotics for such inferences is 'ahductiun'.Abduction is a case of synthetic inferelll.:e '\\ here we find some \ery curiouscircumstances, which would he explained by the supposition that it was a elseof some general rule, and thercupon adopt that supposition' (Eco 1976: 131,citing Pierce ii. 624.). Elsewhere, Eeo writes 'Abduction ... is a tentative andhazardous tracing of a systcm of signific.ltion rules which allow the sign 10acquire its meaning.... [itl occurs with those natural signs which the St'Oiescalled indicative and which are thought to be signs, yct without knowing whatthey signify' (Eeo H)8+ +0). Abduction uwers the gre}' area where semioticinference (of meanings from signs) merges with hypo/helical illfrrtnus of anon-semiotic (or not convemionally semiotic) kind, such as Kepler's inferencefrom the apparent motion of Mars in the night sky, that the planet traxelled inan elliptical p:1Ih:

Abduction is 'induction in [he service of explanation, in \\ hich a new empiril'lll ruleis created 10 render prediclable wh:lt would mhcn\'ise be mysterious'. . Alxluctivn isa variell' of nondemonstrativc inference, oosed on Ihe logiell fallacy of affimling theanteccd~nt from the consequent ('if p then q; but 'I; therefore p'). Gi\en true premiscs,it yiclds cunclusions that arc not necessarily true. Ne\'crthele~~, abduction is an indis­pensable inference principle, bccau~e it is the basic mechanism that makes it p()~~ible

to constrain thc indefinitely largc number of explanariuns compatible with any CYent.

(Boyer J994: 147, citing J. Holland (/ 01. 1986: 8<))

I have a particular reason for using thc terminology of 'indexical signs' and'abductions' therefrom in the presenl connection. 1\0 reasonable person couldsuppose that an-like relations between people and things do not imolve at leastsome form ofscmiosis; howsoc\'er one approaches the subject there seems some­thing irreducibly semiotic about art. On the other hand, I am particularly anxi­ous to avoid (he slightest imput;ltion that (visual) art is 'like language' and that(he rcle\'anr forms of scmiosis are language-like, Discovering the orbits of theplanets is not in rhe least analogous [Q intcrprcting a sentence in an~' naturallanguage. Kepler did not discover the 'grammar' of planetary ~l1ot.lOns, forthere is no equivalent to grammar in nature. On the other hand, SClentlsts oftenspcak (metaphorically) of their dar.\ as 'meaning' this or that, in (~ther wordspermitting certain infercnces which, if they do not appeal to estabhs.hcd.phys­ical laws arc abductions. The usefulnL'Ss of thc concept of abduction IS thatit design~tc.<; <1 class of semiotic inferences which a['e, by definition, wh~lIydistinct from the semiotic inferences we bring 10 bear on the undersmndmgof language, whose 'literal' undcrsl.;lIlding is a matter of obsen'ing semiotic

The Hleor)' oIlIlt Arl No:us

coO\'entions, not entertaining hypotheses derived aJ !IO( from the 'L":tSC' underconsideration (£co IQ8.r -10). Abduction, though a scmiOlic concept (actually,it belongs to logic rather than semiotics) is useful in lhat it functions to setbounds to linguistic scmiosis proper, so that we cease [(I be tcrnptcd to aprl}linguistic models where thcy do not aprl}, while remaining free [(I posit infer­ences ofa non-linguistic kind.

For our purposes, a more perspicuous example of abductivc inference froman index is the instance of smiling 'meaning' friendliness. Vcry much part ofthe theory I am proposing is the idea that we approach art objects (and mem­bers of a larger class of indexcs of agency) as if the)' had 'physiognomies' likepeople. \Vhen we Sec a picture of a smiling person, we ilnribute an attitude offriendliness to 'the person in the pielure' and (if thcre is one) the siuer or 'sub­jel:t' of thc picture. We respond to the picture in this way bL'Cause Ihe appear­ance of smiling triggers a (hedged) infcrenL'C that (unless they are pretending)Ihis person is friendl), just :1.<; a real person's smile would trigger the sameinference. Wc ha\'(;, in short, acccss to 'another mind' in this way, a real mindor a depicted mind, but in either case thc mind of a well-disposed person.Without pausing t'O unravel the very difiicult question as to the naturc of therelationship ben,'ccn real and dcpicted persOrL<;, the point I want to emphasizehere is that the means we generally have (0 form a notion of the disposition andintentions of 'social others' is via :l large number of :lbdUL"(ions from indexeswhich are neither 'semiotic conventions' or 'laws of nature' but somcthingin between. Furthermore, the inferential schemes (abductions) we bring to'indexical signs' arc frequcntly \'cry like, if not actually identical to, the oncswe bring to benr on social others. These may seem \'ery elementary points, burthey are essential to the anthropology of art.

The minimal definition of the (visual) 'art' situ.uion therefore invoh'CS thepresence of sOme index from which abductions (belonging to mam' dilTeremspecies) may he madc. This, by itself is insufficiently resrricti,'c, since it willbe apparcnt that. formal reasoning and linguistic scmiosis apart, the greaterpart of 'thinking' consists ofabductions ofone kind or :lnother. To restrict thescope of the discussion, I propose th:lt the l":tlegory of indexes relevant to ourtheory are thosc which pennit the abduction of 'agency' and specifically 'socialagency'. This excludes instances such as scientific inferences about the orbitsof planets (unless one imagines th:lt Ihe planets arc social agent<;, which ofcoursemany peoplc do). J-1owc\'er, the restriction is narrower than this, and excludesmuch else besides scientific hypothesis-formation. The stipulation I make isthat the iI/de.\' i.l irsdfseel/ (IS fhe vl/f(()/l/e, (flld/vf fhe imfl'llllfe/ll 4; socia! IIge/It')'.

A 'natural sign' like 'smoke' is not scen as the outcome of :my social agen~y,?ut as thc ~lItcome o~ a. natural .causal process, combustion, so, as an indcx of1IS non-SOCial cause, II IS of no Interest 10 us, On Ihe other hand if smoke isseen as the. indc:\ of fire-setting by hUlllan agents (burning sll'idde~s, say) thenthe abductIon ofagency occurs and smoke bL"Comes an arlcfactual index, as well

ICJ 'nIl' Theory of/ht' Art Nt,I'uS

as a 'natural sign', '1'0 give anofher example, lei us suppose fhat, strolli~g alongthe beach, we encounter a SlOne which is chipped in a rather suggestl\'e way.Is it perhaps a prehistoric handa.'(e~ It has become an 'artefact' and hencequalifies for com..ideration. It is a tool, hence, an index of agency; .I~th theagency of its maker and of Ihe man who used It. It may not be ,ery mterest­ing' as a L':lndidate object for theoretical consideration in the 'anthro~>olo~ofart' contexI, bUl it certainly may be o;;aid to possess Ihe minimum quahficatlol15,sincc we have no a priori ffi(,>;\nS ofdistinguishing 'arle('lcts' from 'works of art'(Gelll(}t)6), 'fhis would Ix: true even if I concluded thar the chipped slone wasnot actually made by a prehistoric artisan, but, h;wing taken it home anyway, Idecide to use it as an orllamcnl Jor my mantelpiece. Then it has hecome anindex orm)' agency, and qu,lli!ies yet again (besides which it is now obviouslya 'work of an' i.e, a 'found object'),

2.4- The Social Agent

However as is <Tencrally the case with definitions, the stipulation that the indexmust be "seen a~ the outcome, and/or the instrument of, social agency' is itselfdependem on a still undefined concept, Ihat of 'social agent'-thc one ",?oexercises !>Ocial agency. Of course it is not difficuh fO give examples of SOCialagents and social agency. Any person must be considered a social agent, at least

potentiallr· .Agency is attributable to those persons (and things, see below) who/which

are seen as initiating causal sequences ofa particular type, that is, events C:lLlsedby acts of mind or will or intention, rather than the mere c~l:caten.atio.n.ofphysical events. An agent is one who 'causes cvents to happen 1Il dlclr Vlcm­ity. As a resull of this exen:isc of agency, certain events transpire (not neces­s~ril\' fhe specific evems which were 'inlended' by the agent). Whereas chainsof physica~malerial GlUse-and-eftecr l.:onsisl of 'happeninl?s' which can beexplained by physicdllaws which uhimalely go\ern the um\'ers~ a,s a w?ole,agents initiate 'acfions' which are 'caused' by themselves, by then lIlte.n~lons,

not by the physical laws of the cosmos, An agent is the source, the ongm, ofcausal events, independently of the state of the physkal uni'Tr.se,. .

Aetunllv the nature of rhe relations between the ngellt's behefs, mlenflons,etc. and Ih~ external events he/she causes to happen by 'acting' are philoso­phically very debatable, Philosophen; are far from agree~ as to, thc n,~lture of'minds' harbouring 'intcntions' and the relation between 11111cr 1I1tent'l01lS an~real-world events, Sociologists, also, have every reason to be a,w'lre thar agent,sactions vcrv often have 'unintended consequences' so rhat It eanno~ be saidIhal rcal-w~r1d (social) e"enl'S arc just transcriplion~ of wh~t ager~ts IIltendedto happen. forrunnlely, in order to carryon .thiS p~rt1cular diSCUSSIOn, I. do nothave to solve problems which have preoccupied philosophers for eenf~n,es. Forthe anthropologist, the problem oC 'agency' is nOf tl ma[(er of prcscnbmg the

17

~ost ralio~lal or defensible notion of agency, in thai the anthropologist's task~ to dcsc,nbc forms ,of thoughf which could not sf'Jnd up £0 much philosoph_Ical serulln)' but whIch are none the less, socially and cogniti\-e1y pr:ICficable.

Fo: the ant~ropol,ogist 'folk' notions of agency, extracted from e\'ervdavpra~tlccs and dlscurslye fonns, are of concern, not 'philosophically dcfen;ibl~'~OfIO~S of ag.enC)'. Some p~i'osophers bclie\'e that 'folk' notions aoour agenC)',m~e~tlon, mmd,. etc. constJtulc a set of philosophicJ.lly defensible bclier..., bUIthIS tS of no particular concern to us, I am going to take seriously Ilofions aboutagency whid} eyen these philosophers would probably not want to defend forexample that agency can inhere in graven images, not 10 mentiOIl motor'tars(see below), I do so because, in practice, people do attribute intentiOlls andawareness to objec~ like cars and images of the gods. The idea of agency isa culturally presenbed framework for thinking aboul causation when whathappens is (in some ngue sense) supposed to be in fended in adv~nce b\' someperson;'.lgcnf .or thing-agent. Whene\'er an e\'enr is bcliered to happen becauseof an mtentlon' lodged in the person or thing which initiafCS the causalsequence, that is an instance of 'agency'.

Putting Ihe word 'social' in from of the word 'agent' is in a sense redundantin so far as the word 'agency' primarily serres !O discriminate between 'ha~penings' (caused by physical laws) and 'actions' (caused by prior intenlions),'Prior intentions' implies the altribufion £0 the agent ofa mind akin fO a humanone, if nor identical, Animals and material nbjecis L':ln hn'C minds and intcn­ti(~ns attributed to them, bur these arc always, in some residual sense, humanmmds, because we have access 'from rhe inside' only to hum:m minds, indeedto only one oCthese, our own. Human minds arc inevitably 'social' minds, to

the extcllI that we only know our own minds in a social context of some kind,'Action' CJnnOf really be conceptualized in other than social terms. Moreover,fhe ~nds of agency which arc altribufed to an objects (or indexes of agency)arc mherently and irreducibly social in thaf art objttts never (in any relev­ant way) emerge as agents except in \ery specific social contexts. Art objectsare nOf 'sclf-sufficienr' agents, bUf onl:r 'secondary' agents in conjunction withcerrain specific (human) associates, whose idcmitics I discuss below. Thephilosophical fhoory of'agents' presupposes the aUfOnomy and sdf-sulliciencyof the human agent; bUI I am more concerned with the kind oC second-classagency which artefacts acquire once rhey become enmeshed in a Icxture ofsocial relat'ionships, However, within this relational reXflIre, artefacts can quitewell be lrc.1\ed as agents in a v;lriety of ways.

2,5· 'Things' aJ Social Agel/ts

The immediate 'other' in a social rclal-ionship docs not ha'-e to be another'human being', I\[Y whole argulllenr depends on Ihis not being the L':lSe. Socialagency can be exercised reb,i\·e fO 'things' and social agency Gm be exercised

18 Tilt Thtll'Y of/lit Arl NtX/lS Thr Thtory of/ht Arl Nt.t:us 19

b~ 'things' (and also animals). The concept of social agency has t~ be formu­lated in this WTY pcrmissi,"c manner for empirical as well as theorctIL"3! reasons.It just happens' 10 be patenrly the case that persons form what arc c\,idcntlysocial relations with 'things'. Consider a little girl with her doll. She loves herdoll. Her doll is her best friend (she says). Would she tosS her doll overboardfrom a lifeboat in order to save her bossy elder brother from drowning? Nowav. This rna\' seem a (ri,ial example, and the kinds of relations small girlsfor~ with their dolls aTC far from being 'typical' of human social bch:\\'iouf.Hut it is not a lri,-ial cX:lmple at all; in fact ir is an archetypal instance of thesubject-maner of the anthropology of art. We only think it is not because it isan aITront to our dignity to make comparisons beTween small girls showeringaffection on their dolls and us, maTure souls, admiring Michelangelo's David.BUl what is David if it is noT a big doll for grown-ups? This is not really a mat­ter of dcvaluing D(I'1)id so much .1S revaluing little girls' dolls, which arc Trulyremarkable objecL", all things l:onsidered. They arc cerminly social hcings­'members of the family', for a time at any rate,

From dolls to idols is but a short step, and from idols to Sl:ulptures byMichelangelo another, hardly longer. But I do not wish to confine the notionof 'social relations beTween persons and things' to instances of this order, inwhich the 'thing' is a representation of a human being, as a doll is. The con­cept required here is much broader. The ways in which social agency can beinvested in things, or can emanate from things, arc exceedingly diverse (seeMiller 198j for a theOrCliC'J.1 analysis uf 'objectification').

Take, for instance, the relationship benveen human beings and cars. A car,just as a possession and a mC:lns of transport is not intrinsically ~ ~ocus ofagency, either The owner's agency or its own. But it is i~ fact \ery.dl~cult fora car owner not to regard a car as a body-part, a prostheSIS, somethmg lIl\'esTedwith his (or her) own social agency vis-a-vi.~ other social agents. Just as a sales­man confronts a potenti'll client with his body (his good teeth and well-br.ushe.dhair bodily indexes of business competcnce) so he confronts the buyer wnh hiScar (a :V1o~deo, late registration, black) another, detachable, part of his bodyavailable for inspection and approval. Conversely, an injury suffercd hy the caris a personal blow, an outrage, e"en though the damage can be made good andthe insurance company will pay. N"ot only is the car a locus o.f the owner'sagency, and a conduit through which the agency ofothers (bad dm-~rs, vandak)may aITect him-it is also the locus ofan 'autonomous' ~gency of ItS uwn.

The car docs not JUSt reflect the owner's personhood, It has perso.nhood as acar. For example, I possess a Toynt,1 which I esteem nllher th,m abjectly love,but since To\,otas are 'sensible' and rather dispassionate cars, my Toyom docsnot mind (it is, after all, Japanese-cars have distinct ethnicities). In my fam­ily, this Toyota has:I personal name, Toyolly, or ~Olly' f~r short~ .\ly Toyota. ,. hi d ~.'d·,"''''· it onl)' breaks down III rclat1\'eh' minor ways atIS re 13 e an co...., ........, '.;imes when it 'knows' that no great inconvenience will result. If, God forbid,

my To)ota were to break down in the middle of the night, far from home Ishould consider this an act or gross treachery for which I would hold the ~rpersonally and morally culpable, not myself or the gar.tge mechanics who ser­vicc it. Rationally,} know that such scntiments are somewhat bizarre, but I alsoknow that 99 per celli of car owners anribute personality to their cars in muchthc same way that I do, and that such imaginings contribute LO a satisfact_ory ffWJl1S vivendi in a world or mcchaniC:ll dcvices. In eITeer, this is a form of'religious belicP (\'ehicular animism) which Illceept because it is part of 'carculture'-an important element in the ,Ie fiu:to culture of twenticlh-centunBritain. Because this is a form of 'animism' which I actually and habituall;'practise, there is c"ery reason to make mention of it as a template for imagin­ing forms ofanimism that I do not happen to share, such;ls the worship of idols(see Chapter j below, and particularly Sections j.8-9, where the discussion ofthe 'agency' of im'1ges is taken up in grealer detail).

So, 'things' such as dolls and cars can appear as 'agent.,' in particular socialsituations; and so-we may argue--can '\\orks of art', While some form ofhedged agreement to these propositions would, perhaps, be widely concededin the current climate of conceptual relativism and pragmarism, iT would befacile in the extreme not to obser\'c that unwelcome contradictions arrive intheir wake.

2.5.1. Paradox Elimination

An agent is defined as one who has the L-apaeity to initiate C'dusal e\'ents inhis/her vicinity, \\-hich cannot he aseribed to the current m.te of the physiL-a1cosmos, hur only to a special category of mental states; that is, intentions. It iscontradictory 1'0 assert that 'things' such as dolls and cars can behave as 'agent5'in contexts of hUlll;ln social interactions, since 'things' cannot, by definitionhave intentions, and moreover, such causal c\'ems as uccur in their vicinityarc 'happenings' (produced by physical causes) not 'actions' relenlble to theagency exercised by the thing. The little girl may, possibly, imagine that herdoll is another agent, but we arc obliged to regard this as an crroneous idea. Wec:m preoccupy oursch'es with detecting the eogniti"c and emOTional factorswhich engender such erroncous ideas-but this is 'cry different from propos­ing a theory, as I seem to be bent on doing, which accept5 such palpable errorsin agency-anribUlion as basic postulates. This appears a dangerous courseindeed. A 'sociology uf actiun' prcmised on rhe intentional nature of agency,undermines itself r.ll:llly by imroducing the possibility that 'things' could beagents, because the whole intcrpretative enterprise is fuunded on the strict scp­ar.ttion berween 'agency'---excTeiscd by sentient, cncultuflned, human beings­and the kind of ph) sical C:lUsation which explains the beha,'iour of merc things.Howner, this parado:. c:m be mitigated, initially, in the light of the followingconsiderations.

Whatever happens, human agency is CXCH.:iscd within the material world.Were the kinds of material cause and effect with which we are familiar not inplace, ;menrional action, anion initiated in a social oontexi and with socialobjectives in "iew, would be impossible. We can accept that fhe C,lUsal chainswhich <IfC initiated by intentional agenrs come into being as Slates of mind,and Ihal they are orientated towards the states of mind of social 'others' (i.e.'patients': see bclow)-but unless there is some kind of physical mediation,which alwa~'s does exploit the manifold causal propenies of the ambient phys­ical world (the environment, the human body, ctc.), agent and patient will nOIinteract. Therefore, 'things' with their thing-I), causal properties arc as essen­tiallO the exerl..;se of agency as stales of mind. In fact, it is only because the(alls(lll1/i"~u in the vicinity of an agenr assumes a certain configuration, fromwhich an intention may be abducted, that we recognizc the prC5(.'f1ce ofanotheragent. We recognize agenLJ, a post[acto, in the anomalous configuration of thecausal milieu-but we cannot detect it in advance, that is, we cannot tell thatsomeone is an agent before they act as all uffml, before they disturb the causalmilieu in such a way as can only be attributed to their agency. Because the attri­bution of agency rests on the detection of the effeas of agency in the I.:ausalmilieu, rather than an unmcdiated intuition, it is not paradoxical to understandagency as a factor of the ambience as a whole, a global characteristic of theworld of people and things in which we li\'e, ruther than as an attribute of thehuman psyche, exclusiyely. The linle girl's doll is not a scif-sulTIcienl agentlike :m (idealized) human being, even the girl herself docs not think so. Butthe doll is an emanation or manifestation of agency (actually, primarily thechild's own), a mirror, vchide, or channel ofagency, and hence a SOUft.-c of suchpotent experiencUi of the 'co-presence' of an agenl as to make no difference.

I am prepared to make a distinction between 'primary' agents, thai is, inten­tional beings who are categorically distinguished from 'mere' things or arte­facts, and 'secondary' agents, which arc artefacts, dolls, cars, works of art, etc.through which primary agents distribute their agency in the causal milieu, andthus render their agency effective. Rut to call artefactual agents 'secondary' isnot to concede that they are not agents at all, or agents only 'in a manner ofspeaking'. Take, for instance, the anti-personnel mines which have t:aused som:ln\, deaths ;\lH.1 mutilations in Cambodia in recent years. Pol Pot's soldiers,who·ltlid these mines, were, clearly, the agents responsible for these crimesag;linst innocent peopk. The mines themselves wcre just 'instruments' or'tools' of destruction, not 'agents of destruction' in the scnse we mean whenpinning moral responsibility on Pol Pot's men, who could hal'e acted differ­ently whih:: the mines could /III/help exploding: once troddcn on. It seems sense­less t'O ;lltriburc 'agency' to a merc lethal mechanical device, rather than its

culpable user. . ..Out not so fasl. A soldier is not just a man, but :1 man Wllh a gun, or 10 IhlS

ease wilh a box of mines to sow. The soldier's weapons aTC parIS of him which

1'he 'I'll,,!)/), olille Art Nexus20 The fheQry II[/fle Arl NexlIJ 21

~ake him what he is. We cannot speak of Pol Pot's soldiers without referringIII the same breath, to their weaponrv and the social '·ont"xt 'nd .[. '........, ..... millarytactICs \\'hlCh the possessIOn of such weaponf" implies Pol l'ot".. -,. "men werecapable ofbemg the kmd of(~·ery malign) ab'"Cnts that they were only beeauseof the artef~cts the~ had .af theIr disposal, which, so to speak, turned lhem frommcre Ill~n IntO dcvlls wllh eXlraordinary powcrs. Their kind of agenn \\'ouldbe u~thmkab~e except i.n conjunction with the spatio-temporally e;pandedcapacity for \'lOlence whIch the possession of mines makes possible PIP ,Id'. . 0 Otsso lers possessed (hke all of us) wh.at I shall later discuss as 'distribUted per-~nhood . As agents, they were not ,ust where their bodies were but in man'dlffer~n~ plac.c~ (and times) simultaneously. Those mines wer~ oomponen~~f ~elr Identities as human persons, just as much as their fingerprints or thehtames of hate and fear which inspired their actions.

If we t~ink of an anti-personnel mine, not as a 'tool' made use ofbv a (oon­~ptually lIldepen~en~) 'u~r', but, more realistically, as a component ·of a par­tlcu~ar type of sooalldenmy and agency. then we am more readih' see whva ml.ne can ~ seen as an 'agem'-thar is, but for this artefact, this ;gent (th~soldIer + ~lOe) could not exist. In speaking of artefacts as 'scoondar} agents'I am rcfernng [0 the fact th.:u the origination and manifestation ofagenc\' takesplace in a milieu which consists (in large part) of artefacts and that ~~ntsthus, 'are' and do not merely 'use' the artefacts which conn~cl them to ~ocia;others. Anti-per~nnel min~ are not (primary) agents who initiate happeningsthrou~h a~ts of WIll .for which they are moraUy responsible, granted, but theyare obJect~~e c~bodlments of the power or capacity to will their use, and hencemo~al entitles III themseh·Ui. I describe artef.1cts as 'social agents' not becauseI WIsh to promulgate a form of material-<::ulture mvsticism bUl ani" in \'iew ofthc ~act ~hat obj.ectification in artef.1cl-form is how'social a~cney m;nifests and~eallz~ :tsclf, \'Ia the proliferation of fragments of 'primary' intentional agentsIII theIT secondary' artefactual fonns.

2.5.2. Agents and Patients

~lany more ex~mple~ of social agency being attributed to 'things' will be pro­VIded as the dISCUSSIon proceeds, bur rhere is another issue which needs tobe dealt with in this connection. There is a special feaLUre of the concept of~gency th'Il.1 am ad ~ancing to which I must draw particular attention. 'Agency'IS us.ually dl.~cussed III rchuion to t'he pel'm.:menL dispositional characteristics ofparticular entities: 'here is X, is it, an agent or not?' And the answer is-'thatdepe~ds on whether X has intentions, a mind. awareness, consciousness, etc.'Th~ .Issl~e of'agency'!s thus raised in a classificatory context, classifying all theentltles I~ the world lI~to rhosc that 'COunt' as agents, and those d1at do not.Mosl philosophers bcltel'c that only human beings arc pukka agents, while afew more w.ould add some of th~ mammals, such as chimpanzees, and somewould also IIlcludc computers \l'llh appropriately 'intelligent' software. It is

22 Jlle In~IJ'Y"J""J""'_""_

import3m w emphasize thai J am nol raising the question of 'agency' in any­thing like this 'classificatory' sense. The concept of agency I employ is rela­tional and contcxt-dcpendent, not classificatory and contcxt free. Thus, torevert to the 'car' example; though I would sponranwllsly allributc 'agency' (0

Ill\' car if it broke down in the middle of the night, far from home, with me init ", do not think that my car has goals and intentions, as a "chicular agent, thata:e independent of the usc thaI [ and my family make of my car, wilh "hieh itcall co-operatc or not. My car is a (potential) agent with n··"SIX'Ct to me as a'palienl'. nOI in respect to irsclf. as a car. It is an agent only in.so far as I am apatient, and it is a 'patienl' (the counterpart of an agent) onl~ IS so far as I aman agent with respect to it.

The concept of agency 1employ here is exclusively r<:lational: for any agent,there is a patielll, and conversely, for any patient, there is an agent. Thiswnsiderably reduces the ontological ha\'OC apparently caused by attributingagenc~ freely to non-living things, such as cars. Cars are not human beinb'S, butthey act as agents. and sufTer as patieills 'in the (causal) \-ieinit),' of humanbeings, such as their owners, vambls, and so on, Thus I am not really indulgingin paradox or mysticism in describing, as I shall, a picture painted by an artistas a 'piltiem' with respect t.n his agency as ::an ::artist, or the victim of a cruelcaricature as a 'paticnt' with respect to till: image (agent) which traduces him.Philosophers may rest content with the notion that. in such locutions, the onlypuN/a agents are the human ones, and thaI cars and caricatures (second'lryagents) could ne"er be pukka agent... , I, on the other hand, am concerned notwith the philosophical defmilion ofagency sub Specl~ a~JeTlljf(JljS. I am concernedwith agent/patient relationships in the fleeting contexts and predicaments ofsocial life, during which we certainly do, transactionally spcaking, attributeagenc), to cars, imagL"S, buildings, ;md many other non-li\'ing. non-human,

things.In what follows, we will be conl.:el'1led with 'social agems' who may be per­

sons, things, animals, divinities, in fact, anything at all. All that is stipulated isthat with respect TO (Iny given frill/Sllcrion between 'agents' one agent is cxercis­in'" 'agenc,,' while !.he other is (momentarily) a 'patient'. This follows fromth~ essentially relational, Transiti\'e, and causal implications of our notion of':lgenc)". To be an 'agenr' one must act with respect to the 'patient'; the patientis the object which is causally affected by the agent's action. For the purposesof the theory bein'" dnelopcd here, it will be assumed that in any gi"cn trans­action in w'hich at>gency is manifested, there is a 'patient' who or which isal101h~r 'potential' agent, capablc ofacting as an agent or being a locus aI'agency.This 'agent' is momentarily in the 'patient' position. Thus, in lile 'car' ex­ample just considered, if my car breaks down in the middle of the night, I a~in the 'patient' position and the car is the 'agent'. If I should respond to thiSemergency by shouting at, or maybe even punching or kicking my unfonu~ate

vehide. then I am the agenl and the car is the patient, and so on. The vanous

possibilities and combinations of ageneY/paliem:y will be described in detaillater on,

Jl is import:ml to understand, though. thal 'palicms' in agent/patien! inter­actions arc nOI entirely passive; they may resist, Thc concept of agency impliesthe onreorning of resistance, difficulty. inertia, elc, An objects are chamcter­iSlically 'diffieuh'. They are difficult to mllke, difficult to 'think', difficult rolr.ms.1Ct. Thcy fascinate, compel, and entr.lp as well as delight the spcctator.Thcir peculiarity, intransigence, and oddness is a key factor in rheir CfJiC;lC\' associal instrumelllS. _~1oreover, in the vicinity of art objects. struggk's for c~n­

rrol ::arc played out in which 'patients' intcn'ene in the enchainmenr of inten­tion. instrumenl, lind n.-'Sult, as 'passive agents', that is, intermediaries betweenultimate agents and ultimate patients. Agent/patient relations form nested hier­arch,ics~\'ho.s~ characteristics wil~ be descri~d in due course. The concept ofthe patlCllt IS not, thereforc a slluplc one, III that being a 'patient' rnav be aform of (derivative) agency. '

2,6, The Artist

Howc\·er, we still h.we nOl specifk-li the situation sulliciently to circumscriberhe. scope of an 'anthropolobrica1 theory of art'. Agency can be ascribed La'thlll.gs' witho,ut !.hi~ giving rise to anything panicnlarly recalling the pro­ducnon and CIrculation of 'art'. For this to be the case it seems neccs.~\T\' tospecify the idcnrity of the parricipaills in social relations in the vicinirv of the'inde:\:' f"dther more precisclv. '

The kinds of'inde:\:' with'which the anthropological theory ofart has to dealarc usually (but not always) artefact.~. These artefacts have the capacity to indextheir 'origins' in an act of malllljoallre, Any anef.1ct, by virtue of hci~g a man­ufaclured thing, motivates an abduction which specifies rhe idelllit\' of thc'lgetH who made or originated it. Manufactured objL"t."t.. are 'caused' by theirmakers, just as smoke is caused by fire; hence manufacfUred objects are i~deX(.'S

?f their ~lakers. The indl':\:, as manufacturL'1..I object, is in the 'patient' positionIII a socI::a1 relationship with ilS m:lkcr, who is an agent, and witham whoseagency it would not exist. Since art-making is the kind of making with whichwe are primarily wncerncd, it might be most convenient to L':III the one towhom the authorship of the index (as a physical thing) is attributed, 'the artist'.Where\'er it is appropriate, I shall do so, but it is important to note that rheanthropology of :1ft cannOt be exclusively concerned wirh objects whose exis­renee is attributed to the agency of 'artists', especially 'hulllan' artists. [vIan\,obj~cts which arc in fact art objecL~ m::anuf.1crured by (human) artists, arc n~thcltc\"ed to have ~riginated in lhaT way; they are thought to be of di"ine origintlr to have mystenously made lhemselves. The origins ofart objects can be for­gotten or .co~ccalc<.l, blocking ofT rhe abduction ICllding from the existence ofrhe matcflalllldex La the agency of an :lrtisl.

24 The [{leor)' vOile Art ""exII'> liz" TheM)1 ofthe .1f/ j\,'fXU.1 25

2.7. The Recipient

Art objects lead vcry transa<:tionallives; being 'made by all artist' is only thefirst of these. Often an art' object indexes, primarily, not the moment and agentof its manufacture, but some subscljucnl, purely transactional, 'origin'. Thisapplies, for inSTance, to ceremonial valuables in Melanesia (such as Kula shells)whose actual makers (who are not in the Kula system) are forgotten-Kulashells 'originate' with whoever possessed lhem as a kitoum, that is, as unen­cumbered ceremonial property (Leach and Leach 11)83).

Similarly, in the Victoria and Albert ;\luseum, one may sec the beautifulcan'cd ony'x cup of the Mogul emperor, Shah Jehan. This cup is Shah Jehan'skill/um for all that it is now British government property. But there is a differ­ence, in that in Shah lehan's cup, we sec, first and foremost, the power of the~logul emperor co command the services of craftsmen possessing more skilland inventiveness than any to be found nowadays. Shah lehan's agency is notas a maker, but as a 'patron' ofart, and his cup indexes his glory in this respect,which contemporary potentates can only cmulate in feeble, vulgar, ways.

Thus a second abduction of agency which an index in the form of an artc­fact normally mocivates is the abductiun of its 'destination', its intcnded recep­tion. Artists do nut (usually) make art objects for no reason, they make rhem inorder that they should be seen by a public, and/or acquired by a patron. Just asany art object indexes its origins in the activity of an artist, it also indexes itsre~eption by a public, the public it was primarily made 'for'. A Ferrari sportscar, parked in the street, indexes the class-fraction of'millionaire playboys' forwhom such cars are made. It also indexes the general public who can onlyadmire such vehicles and envv their owners. A work of contemporary artindexes the contemporary art p~blic, who constitute the intended recipients ofsudl work. If the work is to he seen in the Saachi gallery, it indexes this famouscollector and his patronage of contemporary art. And so on. In the coursc oftheir careers, art objects can have many receptions. "Vhile I am able to fecithat I belong (as a gallery-goer and occasional reader of Art Nom and similarperiodicals) to the 'intended' public for contemporary art, I know perfectlywell that the Egyptian art in the British ~luseum was never intended for myeves. This art permits thc vicarious abduction of its original, or intended recep­ti'on, as a compunent of its current, nun-intended reception.

The public or 'recipients' of a work of art (index) are, according to theanthropologic~l theory of art, in a social relation~hip with the index; eithe~ ~s'patients' (in that the index causally affects them III s~me wa~') or as agents III

that but for them, this index would not have come lllto eXistence (they havecau;ed it). The relation between the index ami it~ reee?t.ion will ?e analysedin greater detail in due course. For the present it IS sU~~lent to s~lplllate thatan index has alwavs to be seen in relation to some speclfic receptlOn and rhatthis reception Illay be active or passive, and is likely to be diverse.

2.8. The Proto~}'pe

~o. c.omplcte the. specification of the network of social relationships in theV1Cltllty of art objects, we necd only one more concept, one which need notalways apply, but which very commonly does. iVlost of the literature about 'art'is ~ctually. about representation. That representation is the most complicatedplulosophlcal and conceptual problem stemming from the production and cir­culation of works ofart there is no doubt. Of course, hy no means all 'art' actu­ally is represent:ltional, even in the barest sense, and often it is the case thatthe 'rcpresentational content' of art is trivial, even if the art is representational(e.g. the bottles and guit<lrs in Cubist stilllifes, or the botanically arbitraryflowers and leaves in textile patterns). I do not propose to discuss the problc~

of representation as a philosophical problem in any detail. I should howeverstate that I espouse the anti-Goodman ian view whi~h has been gaini;lg groundrecently (Schier 1986). l do not belie\'e that iconic representation is based onsymbolic 'coll\'ention' (comparable to the 'conventions' which dictate that'dog' means 'canine animal' in English). Goodman, in a well-known philo­sophical treatise (1976), asserts that any given icon, given the appropriate con­ventions for reception, could function as a 'represcntation' of any arbitrarilyselected depicted objecr or 'referent'. The analogy between this propositionand Saussure's well-known postulate of the 'arbitrary nature of the sign' doesnot need to he underlincd. r reject this implausible claim as an overgeneral­ization of linguistic semiotics. On the contrary, and in accordance with thetraditional Yiew, I belieye that iconic representation is based on the actualresemblance in form between depictions and the entities they depict or arebelieved to depict. A picture of an existing thing resembles that thing inenough respects to be rceognized as a depiction or model of i1. A depiction ofan imaginary thing: (a god, for instance) resembles the picture that believersin that god have ill their minds as to the god's appearance, which they havederived from other images of the same god, which this image rescmbles. Thefact that 'thc picture that people have in their minds' of the god's appcar­ance is actually derivcd from their memories of images whicl1 purport ro rep­resent this appearance does not matter. What matters to me is only that peoplebelieve that the causal arrow is orientated in the other wav· thel' believe thatthe gnd, as agent, 'caused' the image (index), as patient, to ;l~su~e a particularappearance.

It is true that some 'representations' arc very schematic, but only very fewvisual features of the emity being depicted need to be present in order tomotivate abductions from the index as to the appearance (in a much morecompletely specified form) of the entity depicted. 'Recognition' on thc basisof very under-specified cues is a well-explored part of the process of visualperception. Under-specificd is not the same as 'not specified at all', or 'purelyconventional'.

26 The Thef!~)1 of/ht: Art "V/'TU:i 27

One c.lrl only speak of rcprescllIarion in visual an where there is rCSClll­hlam:c, triggering recognition. One may need to be TOld that a given index is aniconic rcprcscllliuion of a particular piclorial subject. 'Recognition' m'l)' notoccur spontaneously, but once the necessary information has been supplied,the \'isual n.'cognirion cues must be present, or recognition will still nor occur.

Mc:mwhilc, there arc indexes" hieh refer ro other entities (such as gods,again) which (a) are \isiblc, but which (h) do not permit abductions as 10 thevisual appearance of the enlil~' (god) because rhey lack an) ,"isu:11 recogni­tion cues. Sometimes gods arc 'represented' by stones, bur the god does not'look like' a stone in :mybody's estimation, belic\"Cr or non-believer alikc. Theanthropology ofan has to consider such instances of'aniconic' representalion.as well as the ones in\ol\'ing morc or less overl \"isual cues as TO the app<:aranccof the enrity being reprcsemcd. There are many forms of 'representation' inother Ilords, only one of which is the rcpresentation of visual form. Approx­imately, the aniconic image of the god in the form of a stonc is an index of thegod's spatio-tempor:l.l pTL'Sem:e, but not his appearance. Uur in this <.aSC, thespatial location of the stone is not 'arbitrarily' or 'con\'cntionally' associatedwith the spatiallO<.ation of the god; the stone functions as a 'natural sign' of Ihegud's location just as smoke is a natural sign of the spatial location of fire.

In what follows I shall u.<te lhe term 'the prototype' (of an indcx) to identifythe entity which the index rcpr<.'Scnts vi...ually (as an icon, depiction, CIC,) orn()n-\'isuall~, as in the example just considered. :'IIot all indexes ha\'e pro(~

types or 'reprcscnl' anything distinct from themseh-es. Abstract g<.'omctric pat­terns have no discernible or rele\'ant prototype, but such abstract decorath·cforms arc ofgreal importance theoretiC'dIly, as I shall desaihc lalef. As with theartist (the originator ofan index) and the TL'Cipienl of an index, I hold that therearc \,;Irious types of social agencylpatiency relationships linking indexes andtheir prototypes, where they exist. That is to say, there is a species of agencywhich is abducted from the index, such that thl' prolotype is taken tu he :lll

'agent' in relation W the index (causing it, for instance, to h:n c the appL'al"aneethat it actually has), Conversely, the pl·ototype may be made into a social'patient' \·ia the index (as in 'VOll sorcery', to be described later).

2.9. SUlJIlJ/(IIY

Let me briefly recapituhlte the argument so far. The 'anthropological thcoryof ;)ft' is ,I theory of the social relations that obtain in the neighbourhood ofworks of art, 01' imlcxcs. These social relationships form part of the rebt ionaltcxture of social life within the biographical (anthropological) frame of refer­encc. Sodal rehltions only exist in so (11' as they are made manifest in actions.Performers of SOCi;ll actions :lrc 'agems' and they aet on 'paticnts' (who arcsocial agents in the 'p:llient' position vl~~-ti-'1)is an agem-in-ac1ion). Relationsbetll'een social agcnts and paticms, for the purposes of the amhropological

theory of art, obtain between foul' 'terms' (cntities which can be in relation).These are:

I. Indexes: material entities which motil',lle abductivc inferences, cognitiveinterpretations, etc.;

2. Artists (or other 'origillalOrs'): to whom are ascribed, by abduction,caus.11 responsibility for the existence and characteristics of the index;

3. Recipients: those in relation to whom, by abduction, indexes are con­sidered to exert agency, or who cxert agency \·ia the index;

-I-- Prototypes: entities held, by abduction, to be represented in the index,often by ,'irtuc of visual resemhlance, but not necessarily.

The Art Noms ulld ,ht Indr,t: 29

Table T. Tlu Art Nnus

enp en!

i~'Materialsoun:e of Prototype Rcdpicnl

creative aCI \nhen:nlly COntrOlS "rllsl'~ cause ofdictates to artist anion, aniSI'S actionthe form it appearance of (a, patron)

Artht as a"umcs PCOIQlypewitn""s to imit:lte<l by:leI of artiSI. Realistiecr,-,alion =.Malt'lial Uuff

~ProIOl}'pe

wped by~~01

KKipient th,-,

.uti"., agencyIIS,-,I!: dielat"" lbe C'au..,of Ih~

and intention·self- form t:lten by origination and

ma<Je' index form taten b)'

lnd~", asathe inde-x

...."lhi;;-;'

Appearanre of tmag... or

~I\~~·ipient has

prototype actions of as cau.o;e of

uielate(1 by prototype indexDOwcr over

llCtisl. controlled byth... prototype.

Imaginative art mea"" ofVolt sorcery.

Index. a I(){'"u~ ProtOl)-'peof J>O"'l'r O\"ef" aff""tl'd byprotOI)l'" index

Ke(ipient', Inde>; SOUr«' of ProIocype has

s:repon:ll.' power "'"ef" J'O'M'I" ""ff tIN!

patrOn

dJctated br recipienl. f"f.'C;pient. lmaieanis!'s 5lJII, Recipient as of protOI)l"" U$E'd"il. rnagkal 'Sl'fftator' 10 colltrol artionspowers, etc. submits to of nripi€'nt.Recipient inu ....~. idolatry. Recipient a~

Caplivatc<.l. ,peClaror

Index

ArtiSI

Recipient

Prototype

AGENTr-------~fi-------~

Anist Index Prototype R "'"

PATIENT

~r{;es wer~ bclic\'c~ to .send for sorcerers, to whom they g-.we orders how toshape thCl: .trunks,lnto Idols, and.th~e "cemu" being then insmlled into temple­huts, recel\ cd .thelr prayers and msplred their pril:St.<; and oracles' (Tvlor 1ST:216).. E~·en thts te.~ statement is enough to establish the possibili~' that, irlcertam mstanc.es, It IS ~n agcn~}' in the material of the index, which is held toco~~rol the artist, who IS a p:lllent with respect to this transaction.

I he mor~ common l.:asc is for the material index [Q dictate its form simply~n the ~asl~ of tradition:!l knowledge, rather than by occult instrucliol~

urn{;r,1I1 hiS aCCOllll( ofNdcmb '.. I fin' . , .th' . II lltua s 0 a IctlOn (1968: 72-5), describes~ ~an'lbnlK 0df fig\ltlne~ from th{; wood of the sacred Mukula tree. This tree

S(.'Crdes 00 -rcsemblmg glllll" h . I . . .h d' h" ' m t e ruua {;OnreXf It IS ld{;ntified as 'thes a e W 0 IS causmg menstrual and reproducti\-e lroble' .After being worshipped the If .. . II I ms 10 female patients... ,cclsntuaycutdownad"-d'mlo figunnes n.:sembling babies (u1Ikishi) Th fi' . n Irs ~\OO IS carved, esc gutlllCS aSStst the afflicted

3.2. JlUlex-A~ Artisl-PThe index is the material thing which motivates abdu{;tions of an art-relatedkind. Wh;l! we hav{; to eonsid{;r under this rubric are instances in which themar{;rial index dictates 10 the artist, who responds as 'patient' to its inherentagency. This, of course is Ihe precise inversion of Ihe relationship which wenonnally think of as obtaining between artists and indexes, \\'hich is Arlist-A~ Index-P. However, it is possible, if not \-ery easy, 1'0 find cxamples. Thus,Father Roman }Janc, who wrote an account of the religion of the inhabitants ofthe Anlillt:s, al the behest of Christopher Columbus, reported that: 'Certain

3

3.1. The Table ojAgel1tjPafim[ Relatiolls betmeen FUllr Basic Tums

Where the four terms-Index, Artist, Recipient, and Prototyp<:--cocxist weha\'c, so to speak, the 'canonical' nexus of relations in the neighbourhood ofartobjects, which the anthropology of art must dc.~be and elucidate. BUl, as wewill sec, many instances C-.10 be cited in which 'artists' or 'rttipients' or 'proto­

types' may he I::acking Of only ambiguously present.A theory of the kind being developed here consist.. primarily of a device for

ordering lind dassifying the empirical material with which it deals, rather thanoffering hlW-likc generalizations or prcdktions therefrom. The situations inwhich indexes of an art-like kind can form part of a nexus of social relationsbetween agent.'i are very diverse indeed, and it is necessary to classify them,before orrering oommenurics which will, in the nature of things, only apply to

cerrain of the situations under consideration, and nOI to others. One COlweni~ent approach to the problem of classification is the construction of a table ofcombinations, such as the onc 1 will introdUl.:c at this painl. This table is basedon the premiss that :111 four of the 'terms' so far distinguished can be con­sidered as social agents of dirrerent kinds, and as such, are capable ofbcing inthe 'agent' or 'patient' position v;s-a-l"j$ one another (and in relation to them­selves). Table 1 therefore opposes indexes, artists, prototypes, and recipientsas, respecrivcly, 'agents' (horizontally, reading downwards) and as 'patients'

(vertically, reading across),Tuming LO Table I, I shall now embark on an account of agent/patient rela­

tions between opposed terms, using the suffi..'\es -A and -P to indicate agent andpatient statuS rcspecr.i\'ely. I consider first the index, in the 'agent' position.

The Art Nexus and the Index

30 n/(~ Ar' NtXIfS «lid the 1m/a "flIt' Arllv'exus tllIt/llte JIIt/e.l· ]I

women in regaining their fertility, via further ritual procedures. In the contextofrhe ritual, it is dear thaI the tree which is \mrshipped in its living form andthe canoed 'babies' that arc subsequemly made orils 'Iood, arc both altcrnati,"cforms orlhe fertile/dcsrrucri\'c Mukula tree. It is the idemity orlhe index in i1.<;

living fonn which imposes form on the index in iL'i subsequent, can"ed, stare,the actual clr\·ing being the mere extraction of this inherent (onn. Perhapsan.1logous instances arc not SO hard to disco,"cr in the Westem an doctrine of'truth to materials', the idea that it behon:s the artist (or crafL"man, architect,etc.) to make from his materials what 'they' .....ant, rather than what he wants.The act ofcarving, famously in Ihe instance of Michelangelo's 'sla\'es', is oftenseen as a matter of'libeTating' forms which inhere in Ihe uncul stone or wood.

In this connection, one may also note the more common instances in whichthe artist docs not so much make, as 'recognize' the index. The idea of the'found object' or the 'ready made' place the artist in rhe patient position withrespect to the index, in that it is the index which inherently possesses thecharacteriSTics which motivate its selection by the artist :\s an art object, towhich, none the less his name is attal.:hl.:d as uriginator. Orient:11 :lfl, especiallyJapanese, is particularly rich in the use of found objects fol' artistil.: purpuses,for instance, tlle naTUral boulders whil.:h arc deployed in Japanese gardens.These unC:lrved stones are, by common consent, among The must entranl.:ingobjects 10 he seen anywhere, and the Buddhist g'Jrden-designers who made uscoflhem are far more revered Ihan any stone-caT\"Crs from japan. The \Vest hasan activist notion of artistic creativity, whereas rhe orielllal al'l public esleemsfar more Ihose 'quietist' modcs ofcrcati\'lty in which success attends those whoopen themsekes mosr to the inherent physiognomic apfX-a! of nalUral objects.The same tradition exists in Indian tantric art, which involvcs the cult of nat­urJI, or slightly modified, ri"er boulders, nuts of the coco-de-mer, and so on,as /it/gam and yuni, though in these instances no artist's name is attached to

Ihe object, as it is in Jap:lIl. 'I'he Indian examples really come under the head­ing of 'self-made' indexes, which will be discussed bIer under the heading ofIndex-A~ Index-P.

Modem Western arlists using found objects, most notably Duchamp, arcoslensibly less passive. Ouchamp claimed thaI his ready-mades, such as thesnow-shovel, the bouie-nick, etc., possessed 'the beauty of indifference', thatis, they were selccted on the grounds that nobody could possibly imagine thatDuch:lmp had any particubr rc.lson to select them, rather than something else(which was, in fact, far from bcing the case). Their selection was prcsented asa plll'e act of wi11 on his part, an flat! gm/llilt! in the manner of Gide's amoral­ist hero, Lafcadio Wluiki. Ilo\\'eyer, having 'no reason' to sclcl:t somethingas an object of ready-made art, is of course, a reason, since it is motivated bythe need to avoid seleeling anything for whose selection some reason might beproposed. This was Duchamp's conceil. Consequently, c\'en the purportedly':u'bitrary' ready-mades of DUl:hamp, forced themselves on the artisl (as patienl)

who responded to the appeal of their arbitrariness ,md anonymity, just as theBuddhist landscape artisl'i responded 10 their mutely speaking houlders.

3.3. Jlldex-A~ Recipietll-P

This is the elcmelllary formula for 'passi\'c sJX-'CI<l.rurship·, \\-hich is of coursenot at all difficult to conceptualize or exemplify. Whoever allows his or herattention to be attrJctcd to an index, and submits 10 its power, appeal, or fas­cination, is a patient, responding to the agenc~ inhcrent in the index. Thisagency may be physical, spiritual, polirjcal, elc. as well as 'acsthetic'. The war­rior's shield (Fig. I.zll) is an index which, in contc.\.(. possesses agency, hav­ing the po\\'er 10 demoralize Ihe enemy warriur (the recipient/patient). II wouldbe unwise 1'0 impose any theoretical rcstriction as to Ihe t) pe of agency whichrhe index can mecliate, since, as we will sec, these ma} be exceedingly diverse.~one the less, one may, under the rubric Index-A~Recipicnt~P, ofTcr cer­tain gencral points ~lbout the effect of the index on the rCl.:ipient.

Reconsider, for .1 moment the Asmat shield which terrifies the opposingwarriur. It is surely noTeworthy that thcse designs produce terror by makingterror manifest-these designs seem to have been romp(/S{:d in a mood of terror,and we arc terrified by them (or can easily imagine being terrified by them)because, submitting to their fascination, we are obliged \0 share in Ihe emo­tion which they objectify. The tiger which is about to pounce and dcvour it..\'ictim looks, alxl\'C all, terrijied-of itself, as it wcn..'-:lnd the same is rrue of\nlrriors bearing down with grimaces of fear and rage. The Asmat shield is aftlse /IIirmr, \\hich seems to show the "ictim his own lerroI', when in facl, it isanother's-and in this \\"a) persoades him that he is terrified. Like the famousIrompr-I'tril image (by P<l.mligiani~() of the J\tedusa's hL"ad in Ihe mirror ofPerseus (in the Uffizi g'Jllery) the shield terrifif.:S b~ persuading us that we arewhat il shows.

The same '(lise mirror' effect is obscT\'3hlc in Ill) ri:ld other contexl.., andmay, according 10 Rcnjamin (sec Taussig H)<J3; Iknjamin 1933), consritute thevery secret ofmimesis; that is, to penxi\"e (to intern:llize) is to imitale, and thuswe become (ami produce) what we perceive (see below, Sect. 7.2). 1\1 any rate,Ihe spcctade of a p~linted saint at prayer is conducive 10 piety, of an amorouscouple, to lustfulness, :llld so on, as has been endlessly commented on sinceancient times. Without dilating any further on this venerable theme, we maysuggest that the primary means through which the index :lffects the recipienlis by subverling lhe recipienT'S sellse ofse(f-possessioll in some way. It may bethaT there is, as in the Asmal-shicld case, a spontaneous convergence betweenthe indcx and recipient such th'll the recipient takes un the nature of the index;but usually the medi:l\ion is much more roundabout, as will be described in duecourse. ln~exes can \\ol:k by alienating the spe(;lator as \Idl as by producingidentification (e.g. H) aClIlthe Rig-dud's picture of I.ollis XIV, discussed below,

3·4· Index-A~ Prototype-P

Ilere the index behaves as an agcnt with respect 10 its protot} pc. A familiarexample of this is provided by Wilde's shon slOry The Port",it ofDorian Grey,in which the ageing undergone by thc picture in thc auic causes the prototypeto retain his youthful good looks indefinitely. The fact that this is a fictionalexample need nor deter one from citing it, since anthropology has to deal withfictions as much as with real situations, the two often being hard to Tcll ,lpart,anyway. Another rype of instance of Indcx-A ------t Prototype-P is in sorccry,where the injury done 1'0 a representation of the victim causcs injury to the vic­tim; this vcry common type of image-sorcery ('volt sorccry') will phI)' <ll1

important part in the argument later on, However, as with Index-A ----7

Rccipiclll-P, !ndcx-A ----7 Prototypc-P is morc commonly encoul1lcreJ withrecipients or ,lrtists or sorcercrs in the 'agelll" position as wcll, in a lhrce- orfour-place expression.

Next, I comidcr the 'Index' in the 'Parit:nt' posilion.

which certainly is 11M intended to make the spectator feel 'regal' Of to identifywith the monarch, but to have exactly rhe reverse effect).

Nor can the formula Index-A ~ Recipienl-P be restricted to thoseOOlllexts in which the recipient is confined to sed/Ig the index, ,IS opposed tointerJcting with the index in some other way. For instance, kissing II holy iconwill, some believers hope, dicit the agency of the image in relieving illncs.<>or poveny. Not :111 images with the same ostensible 'reference' 3re equallyefficacious in this rCg;lrdj only some imagt-"S, for example, of Ihe Virgin, ha\'cthis quality. Worshipping the others may assist in one's salvation, but will notcure one's rheumatism. So il is the inherent agency of the material index. .r.l.lher Ihan the Virgin, which is at issue (wha(ever the priest says). Whcrc\'erimages have to be tot/cheLl, Tather (han merely looked at, there is an imputationthat there is inherent agency in thc material index, \\hich is not 10 say that theagency of the prototype is excluded in (hese instanccs.

And herc wc eneounrer a difficulty in presentation. The formula for passi\csp<X;tatorship, Index-A ----7 Recipient-P is so fundamcntal and gcncTaI thatit is difficult lO cite a 'pure' case. Thc 'holy icon of the Virgin which curesrheumatism' is probably more accurately represented as llPrOlOtYPC-i\] --+Index-Aj ----7 Recipient-P, than by the simple formula (the Protorype beingthe Virgin Mary). I intend to deal with three-place and four-place express,ionslater, howe\'cr, where I .....ill explain the significancc of the bT:lckets and ptovidea more perspicuous graphic representation. For the present, it will be sufficientto slipulate that where the indcx is nor seen primarily as the outcomc of anexternal artist's agency, and where it also has no prorol)'pe, it~ agcnq' withrespect (0 the recipient will be a pure l..":Ise of lndex-A ----7 Rccipient-P.

, On 'I>~llisli~' 'CTi"il). SC~ below on Arl;sl-A ---t Artist_P.

JJnle Ar' N<,~'IIS Imd Ihe Judex

3·5· Artist-A~ IlIdcx-P

This is the e1cm~n.tar} formula for 'patronage' and/or 'the spectator as agcnt'.In so far as a reClplCnl can abduct hisfher own agencY from an indcx this for­mula is satisfied. One does not have to lift a finger in ·order to feel th~t one has'~la~e' son.leth.ing. One may readily conceive that a great king (such as LouisX~V.strolhng III the grounds of Versailles) sun-eying the works he has com­mISSIOned and financed, regards himself as the author of rhc scene before hiseyes, fo.r all that ~hese works h:l\e been crc.1ted, in the material sense, by hostsof architects, arusts, craftsmen, masons, gardeners, and other labourers. Thcpatron as provider of the commission is an cfficient cause of the index' hisglorification is its final cause. The patron is thc conduit of the social caus..~tionof such works of art; his agency is therefote readilv abducted from it.

But it is not just great parrons such as Louis XlV who have to be considcrcdin thi~ conn~tion. 'nlere is a morc geneml SCnse, which has been given particularpr?nunence III contemporary aesthetic theory, in which recipicncy as spcctator­ship ~once~ls a form of agency. 'Seeing' is a form of agency in psychologicaltheones of pcrception which emphasize the way in which perception 'goesbeyond the infonmtion gi\'cn'. AceorJing (Q such theories, the mind of theperc~iv~r.activcly 'constructs' the pcrccptu:II image of the thing perceived.Semlotlc{lllrel'pretati\'e theories oran give prominence to the facr fhat what aperso.n sees in .1 picture, or, even more, gleans from an unerance or a text, is afunctlO.n of their previous experience, their mind-sel, their culture, etc. Rcaders,aceordmg to SOI11C critical theorists, h.l\'C been promoted to a status hardlv dis­tinguishable from Ihat oncc occupied by writers; and 1 think thar gallery-'goers

3·6. Recipient-A~ I"dex-P

This is The.elcmentary formula for artistic agency. The index usually motivates~he abductIon of the agency of the person who made it. Thc indcx is, in theseIllsta~ccs, a congealed 'trace' of the artist's creari\'c performance. J\1.uch st­Renalssancc Western art project~ the artist's 'Igene,' in a Yen' ",I'cn' rn po. . _ ... , anncr.The brushwork III works: by Van Gogh emanates an almost p,lpablc ~~n· fh'" ...... seo[ e artl~r ~ p~~nc~, ~me.1rrng ~nd dabbing the still viscous oil paint. JacksonPolloc~ s dnp p:unllngs pro\'ldc even more striking examplcs. Thcy haveno subJI..'Ct ~t all except the ~genC}r of Jackson Pollock himself; rhcy arc (non­reprcsentatl~nal)self-porrralts of a man in frenzied ballis(ic acrivilJ'.' Amonthe very c,arhcst examples of art ofany kind arc the famous hand-prinTS whic~occur beside the cave paimings of I.aseaux, Ahamira, etc. These are particu­larly 'purc' cases of Arri~,-A ----+ Index-P.

'l'hl: Art Nexus lind the index32

34 ffle Art /\'exus lind Jhe 1m/ex The Art A'exu.I and the {/Ide>. 35

have shared, though to .1 lesser degree, in the transfer of agency from the ori­ginators of works of art to the recipients of works of art. I do not wish to

discuss literary theory, since I am only interested in visual art. But one canhardi\' rail to take note of the fact that many members of the contemporaryart p~lblic have actually internalized the view critics rake of their agency asrecipients of art, that is, they attribute creativity ro themselves as spectarors,who cnn 'make something' out of the raw material presellted to them in the artgallen' in effect Recipient-A~ Index-P. Artists collude in this, disclaim­ing th"e'ir own undivided agency ami transferring partial responsibility for their

art to its public.The ideological congruence between the 'speetnwr as agent' theory ami

other aspects of \Vestern individualism is too obvious to need underlining.Gal1ery-goers, who arc mostly middle class and educated, are involved in life­projects which arc predicated on individual freedom, autonomy, ~erso~alresponsibility, and so on. They are hardly likely to abandon these eXlst~ntlnl

attitudes on emering a gallery. They do 110t feel passi\'e; after all, entermg ag-,lllerv is something they do voluntarily, out of motives which can certainly beattrib~ted to their own social agcncy.' In ages past, the art public more resem­bled, in its own estimation, the religious devotees who humbly submit to thepower of the icon and who find causes for personal satisfaction in their verypassivity before it (Index-A ----') Recipient-P). But it is certainly true, now,that spectatorship is seen as a form of agency, even though the spectator rolesimultaneously involves the passive registering of a 'gi ven' art object.

Such considerations aside, there is almost always a sense in which the recip­ients of a work of art can see their own agency in the index. Even if one is not'the patron' WllO caused the work of art to be made, any spectator may inferthat, in a lllore general sense, the work of art was made for him or her. A reli­gious congreg-ation, for instance, is entitled to think that their piety and devo­tion were contributorv to the causation of the cathedral in which they worship,even though this calh~dral was constructed centuries earlier, because they (notunreasonably) believe that the cathedral was erected milh them ill mind, thefuture generations of worshippers therein. They are, in other words, the teleolog­ical cause of the cathedral. Alternatively, in that gallery :Irt is a commodity,gallery-gocrs as consumcrs can infer that thcir 'demand' for art is the fac~orultimately responsible for its existence, just as the existence of any commodityon the market is an index of consumer demand for it.

• On the olhcr hand, c~.ctl\' Ihe samC degrce ofogenc)" could be 'I!filmtcd to thc lhmL,"lnds of pi1­~ims nockin" to .'lIce", to ci~cull1,mbub!cand kiss rhe Koabo, who ore also v"lurll~ry pilgflll1S seek­ing ""lvolion ~throLlgh !heir own cffort" "01 rdigioLl' ou,omot., Thc extent 10 whIch the' cultmaU)"inculcoted belief in ogency of ,pcctotorship is 0 ,creen, concealing from modem souls the extC11l towhieh lheir .crion, .rc dri;'cn hy soci"l imperarive" connot be determined herc: .For our pUl"J"'''''; ",hali, impo"tant are the bclietS thot peuple huld. not whcther thc",e bdids ore IU'lll1ed.

3.7. P}'(!{olype-A ----7 fndex-P

The agency oflhe prototype can frequently be abducted from the index. Therearc ohvioLls ways in which 'prototypes' can hnve agency anributed to them. Inour own an system this kind ofagency is everywhere manifest, since it is essen­tial to the notion of 'realistic representation'. Let me give an example. \Vho isresponsihle for the appearance of rhe Duke of \Vellington in Goya's well­known portrait? We might be tempted to invoke Goya's name here, becausewe are so inclined to impute primary agency to artists with rcspect to worksof art, but this is not so self-evident :IS it appears. Goya's task was to make arepresentation of the Duke of \Vellington, the prototype of the index he prod­uced. He could not produce a picture of a little girl with golden curls andtell the world that this represented the Duke; he would h:lve been regardedas insane and the Duke would have been understandably displeased. In thecircumstances, he had to produce a portrait depicting the features actually pos­sessed by the Duke and regarded as characteristic of his persona, his Romannose, serious demeanour, military attire, etc. It is reasonable to attribute agencyto the Duke of \Vellington with respect to his portrait hy Goya, not becausehe wielded Goya's brushes, but because, in the social nexus which exisredbetween Goya, his painting, and the Duke, he dictated the strokes Goya !tadto make with his brush, merely hy possessing certain features which it wasGoya's task to represent. The Duke, in other words, played a causal role withrespect to the appearance of his portrait. However, this is not a 'pure case' ofPrototype-A ----') Index-P in that Goya's medialion, as artist, is an essentialfeature oftll[ situation: llDukefProtorype-A] ------t Goya/Artist-A]~ Index­P would be a more accurate formula.

Actually, in this instance, the Duke of \\"ellington (as the possessor of agiven physiognomy) is best considered a 'secondary agent', part of the causalmilieu surrounding and permitting the manifestation of the 'primary' agencyof the artist, Goya.

Photography was, Ollee upon a time, considered to be an 'artist-less' mode ofimage-production, and is still so seen by some. The image which forms itselfout oflight emanating from the prototype provides n model for the 'pure' case.

3.8. The Ccntra/iZv olthe Index

These arc all the ahductions which can be drawn from the index as agent andas patient. Turning to Table I, we sec that there are quite a number of otheragent-patient relationships to be considered. Thus we h:lve Artist-A ----')Recipiem-P, Artist-A ----') Prorotype-P, Recipient-A~ Prototype-P andso on. The relevant cells in Table I provide indications of the relationshipsinvolved. However, thcre is a theoretical problem associated with discussing

36 ntr Art Nam ,lIId lit( lndrx

them separately, one by one..>\ basic consTraint on the thL'()r~ being dc\"c1opedhere is that unlc.'is there is an il1Ja, there can be no abductions ofagency, andsince the lopic ohhis theoretical enrerprise is pn..>ciscl) the abduction ofagencyfrom indexes, the index has to be present for analysis to proceed. One C.1I1 con­struct formulae which lack the anist, or the rCl:ipient, or the prototype, but notones which lack the index. It follows that ;} formula such as Artist-A -------t

Recipicnt-P is always implicitly [[Anisr-A] --+ Index-A] -----7 Recipient-P, orsome \'ariant thereof, including the index. So cells in Table I which showbinary relations between terms neither of which is an index arc not theoretic­ally 'well-formed' expressions.

Howc\'cr, for purpos(.'s of general guid:mce, I will hiler gi\-e a brief accountof these illegitimately formed expressions, since there is no point in beingpedantic. Then I will proceed to discuss the 'self-reciprol.:al' relations, Indcx­A ~ Index-P, Artist-j\ ~ Artist-P, Recipicnt-A ~ Rccipient-P,Pr01Otypc-A~ Prototype-P, of which the last three arc not 'well formed'either. I particularly need to do so because the reader ma) be at a loss 10 imag­ine what may be intended by self-reciprocal rclations, for example, an artistbeing an agent with respect to himself.

].8.1. The Logic of 'Primary' and 'Secondary' Agents and Patients

Before turning to the 'illegitimaTe' expressions, I can specify the 'logic' of well­formed expressions rather more precisely; the centrality of the 'index' is not allthere is to it. Let me return to the distinction, sketched in earlier, betwccn 'prim­ary' agents (entities endowed with the capacity to initiate aetioru4ewllts throughwill or intemion) and 'secondary' agents, entities not endowed with will or inten­tion by themseh'es but L"ssenrial to the formation, appearance, or manifestationof intentional actinns. It will he apparent th;1I 'indexes' are, normally, 'second­ary' agents in this scene; lhey borrow their ag-ency from some external source,which they mediate and transfer 10 the patient. It will be equally apparent that'artists' are normally 'primary' agents. They initiate actions on their ownbehalf. This is true e\"en if, as is often the case, they act under the direction ofpatrons. The artist may be a socially subordinate agent, a hired hand, butunless the artist wills it, the index he has been hired to make will ne\"er comeinto existence. In other words, 'subordinate' (but still 'primary') agency of thiskind, is logit"ally quite distinct from 'sL"condary' agency, :IS I have just defined it.

Approximately, then, artists arc primary agents and indexes arc secondaryagents. \Vhat of recipicnts and prototypes? Recipients arc just like artisls; they areprimary agents andlor prim:lry patients, the sources, prime movers, or intended(social) targets of art-medialed agenC)'. Unl~ recipients were primary agents,the art nexuS would not be (as il is) a series of social transactions between per­sons, but a recondite type ofL"ausal interaction be{\\een things.

Prototypes are morc ambiguously situatcd. In general, the prototype of anindex is nOI an intentional or 'primary' agent; an apple ($ay) dOl:s not 'intend'

to appear to us (or to a painter) as round, red :md gT(.'Cn, of certain dimensionsetc: '. or docs it intend to be reprcsemed al all. It just has these \'isual eharac~tensllt.:s, an~ th~ are pa~t of the 'cau5.11 milieu' which the artist exploits, andcontends with, In producmg (as a primary agent) an index which will trigg', "'.' crrecoglllt.lOn as.a ~cpresema,tlOn o~ a!l apple. The apple is a 'secondary' agent~hough, ~n that It 1~ only hy submmlllg' to the apple, by allowing the apple toIm~ress Itself on hl1ll, that the artist can att,lin his goal (as an agent) or'repres_entmg an apple'. Anyone who has tried and failed to draw a 'difficult' objects.uch as the human hand, will know what it is to be an artist in the 'patient' posi~non, confronted by the prototype-as-agenl.

Howe~'~r, prototypes are nOI always just 'secondary' agents, part of thecau.SoII lluhell of art-making :md circulation. Some enritiL'S, unlike apples, 'will'theIr appearance :IS intentional beings, and hence also will their appearanceas subjects for portrayal. Anyone can sec rhat the Louis XIV of HyacintheRigaud's well-known portrait is looking as he wishes to look-he has 'de\"Otedhis whole career, it SL-cms, to perfecting his expression of hauteur and his yervfeatures haw been moulded into a mask of power, a... iflhey were ~f1atcx rathc~than living flesh. His pose and magnificent dress are equallv manifestationsof his royal power of command over appearances, espcciall~ his own. In aninstance such ~s this, th~ distinction between patron (rccipie~t) and prototypethr~aTcn5 ~o dIssolve. Rlgaud's agency, though still indisputably present and'pnmary' IS utterly subordinatcd l'O Louis XI V's as p:lt'ron of the art-makingprocess and also as the one who has the power to appe,1r precisely as he wishesto ~prear (like a god). Here artist and patron/prototype jointly exercise 'prim­ary agency, whereas in the 'apple' case, the artist's agency is primary and theprototype's is sLumdary. In short, where the prototype is 'an objC(:1 n'ot norm­ally thought npable of exercising primary agency 'in the world', then as thesubject of represenlation, it will convey only secondary agency; but where theprototype of an index is an entity {such ,IS a king, magician, divine being, ere.}endowed with the ability to itHcnd its own ;lppearanee, then the !lI'OTotype maybe partly or wholly a primary :lgenl as well as a secondary agent.

l'\ow a word abour the general logical formaT of the expressions I shall bedeveloping in later sections. The pimt of the art nexus is always the index. The~nde~,.h?weyer, i~ never, or at least rarely, a 'primary' agent '(or patient). The!Ddex IS Ilist the 'disturbance' in Ihe causal milieu which re\ C',lls and potentiatesagency L"xercised and patient-hood surTered on either side of i't-by the prim~ary agents, hy recipients (pat rOilS and spectators), b} ,mists, and to a lesserc.xtent, prOtotypes. The index is ,1rticu1:l1ed in rhe callsal milieu, whereas inten­tIOnal agency and patient-hood somehow lie just oUlside it. The index is atonce a prosthesis, an extra limb, of the patron andlor aTlisl, while it is also Ihehandle, anached to the patient-recipient, \\ hich is grasped and manipulated b\"enefllal agents like these. .

I pro\ide a general depiction of the situ:Jtion in Fig. 3.8.1/1, which showsthe index as the region in Ihe causal milieu in which the ;sphl:Te of action' of the

36 Tilt Art NtXIIS Qwl tilt lwlu Tkt Art Nexus and tilt flll/t.r 37them separately, onc by one. t\ basic eonstr-Jinr on the thl.'Or~ being ue\'e1opeuhere is that unless there is:lIl index, there em he no abductions of agency, andsince the topic of this theoretical enterprise is precisely lhe abu uClioll of agencyfrom inuexes, the Index has to be present for analysis ro proceed. One can con­STruct formulae which lack the artist, or the recipient, or the prototype, but notones which lack the index. II follows that a fomlUla such as Anist-A -------+Rccipient-P is always implicitly [lArtist-AJ~ Index-A] -------+ Ikcipient-P, orsome variant thereof, including the index. So cells in Table I which showhinary relalions het ween terms neither of which is an index arc not theoretic­ally 'well-formed' expressions.

Ilowe\ er, for purposes of general guidance, I \yill latcr give a brief al..'(.:ounlof these illegi£imatcly fonned exprt.'s!;ions. since there is no point in beingpedantic. Then I will proceed to discuss the 'self-reciprocal' relations, Index­A ------; Index-P, Artist-A ~ Artist-P, Recipient-A ~ Recipient-P,Prototype-A -----t Prototype-P, of which the last three arc not 'well formed'either. I particularly need to do so bee:tuse the reader Illay be at a loss to imag­ine what may be intLllded by seif-reciprOC:ll relations, for example, an artistbeing an agent with respect to himself.

3.8.,. The T,ogic of 'Primary' and 'Secondary' Agents and Patien£s

Before turning to the 'illegirimate' exprL'Ssions, I can specify the 'logic' of well­fonnl..-d exprcs.~"ions rather more precisely; thc centrality of the 'inde.'+:' is not allthere is to it. Let me return to the distinction, s"etched in earlier, between 'prim­ary' agents (entities endowed with the capacity to initi;lte actiuns/events throughwill or intcntion) and 'secondary' agents, entities not endowed with will or inten­tion by themseh:es but essential to the formation, appearance, or manifestationof intentional actions. It will be apparent that 'indcxcs' are, normally, 'second­ary' agents in this S<.'Cne; they borrow their agency from some external M)uree,whieh they lllL-diate and transfer to the patient. It \\ ill be equally apparent that'artists' arc normally 'primary' agents. They initi:lt'e actions un their ownbehalf. This is true even if, as is often the case, they act under the direction ofpatrons. The artist may be a socially subordinate agelH, a hired hand, bUIunless the artist wills it, the index he ha..o;; been hired to ma"e will ne\'er comeinto existence. In Olher words, 'subordinate' (but still 'primary') agency of thi...kind, is logically quile distinct from 'secondary' agency, :IS I have just defined it.

Approximately, then, artists are primary agents and indcxes arc secondaryagents. Whal of recipients and prototypes? Recipients arc just like artists; they areprimary agenL" and/or primary patients, the sources, prime mO\'ers, or intended(social) targets ofarl-ml..-diated agency, Unless rL'Cipicnts \\'cre primar~ agents,the art nexus would not he (as it is) a series of social tr,msactions betwl..'Cn per­sons, hut a re(;ondite type of causal interaction ben\'een things.

Prototypes are more ambiguously sIIUMcd. In general, the prototype of anindex is nOl an intentional or 'prim:!!'y' agcnl; an apple (say) docs not' 'intend'

to ap~'ar to 1I~ (~r to a paimer) as round, R-d and green, of L"Crtain dimensions,etc. 1'\01' docs It mtend to be repn..."'>Cnted at all. It just has these visual chamc_teristic.:s, and These arc part of the 'causalmillcu' which the artist exploits andconten(~s. with, in producing (as a primary agent) an index which will trigger'recogm~lOn' as.:1 :cpresemation of an apple. The apple is a 'secondary' agent~hough, ~n that It I~ only by 'submitting' to the apple, b~ allowing the apple tolm~res... Itself on him, that the artist e:tn anain his goal (as an agcnl) of'rcpres­cmlllg an apple'. Anyone whn has tried :md f:lilcd to draw a 'diflicult' objectsuch ,IS the hUlllan hand, will know what it is ro be an artist in the 'patient' posi~lion, confronted by the prototype-as-agcm.

Howe\'er, prototypes are not always just 'secondary' agcnt.., part of thccausal milieu ofart-making and circul:uion. Some entilies, unli"c apples, 'will'their a.ppearance as intenrional beings, and hence also will their appearanceas subJects fill' portrayal. Anyone can see that the Louis XIV of HyacimheRigaud's well-known porTraiT is looking as he wishes to look-he has dC\'Ot'cdhis whole e:treer, it seems, to perfecting his expression of hauteur, and his veryfeatures have lx.'en moulded into a mask of power, as if they were of latex ratherthan living flesh. His pose and magnificent dress are equallv manifest':uionsof his ropl power of command over appearances, csp(.'eialh; his own. In anInstance such as this, thc distinction between patron (reciple~t) and protot)'pelhreatens to dissolve. Rigaud's agency, though still indisputably present and'primary' is uucrly subordinated to Louis xrv's as patron of the art-makingprocess and also a.... the one who has the power to appear precisely as he wishesto appe-Jr (Ii"e a god). Here artist and patron/prototype jointly exercise 'prim­ary' agency. whereas in Ihe 'apple' case, the artist's agency is primary and theprototype's is secondary. In shon, whcre the prototype is :In object not norm­ally thought capable of exercising primary agency 'in the world', then as thcsubject of rcprc..<;cntation, it will cOl1\'ey only sC(.'ondary agency; but where theprototype of an index is an entity (such as a king, magician, divine being, etc.)endowed with the ability to intend its own appearance, then the prototype mayhe partly or wholly a primary agent as well as a secondary agent.

Now a word about the gener:ll logical formal of the exprcssions 1 shall bedC\'eloping in later sections. The pivot of the an nexus is alwavs the index. Theindex, however, is never, or at least rarely, :1 'primary' agent'(or patient). Theindex is just the 'disturbance' in the C':lusal milieu which re\'eals and pOlenriatesagency exercised and patient-hood suffered on either side of i't-by the prim~ary agents, by recipients (patrons and spectators), by anists, and to a lesserextent, prototypcs. The index is articulated in the causal milieu, whereas inten­tional agency and patienr-hood somehow lie JUSt outside i1. The index is <Itonce a prosthesis, an extra limb, of the palTon and/or artist, while it is also thehandle, attached to the patienr-recipicnt, which is grasped and manipulated b\external agcnts like these. .

I provide a general depiction of Ihe situation in Fig. 3.8.1/1, which shows1he index as Ihe region in thc causal milieu ill which the 'sphcre of action' of the

38 The Arl Nexus IIlld the 1/11/1",1 The Art NexlIs lind Ihe Indo: .)1)

3.9. The 'lltegitimate' Expressions

1. Artist-A --------; Prototype-Po This is the general formula for 'imaginary'ima'ces made bv artists. From our point of view, an index is an instance ofo ~

'imaginary' image-making, when its appearance is held to have been dictated

primary agent and the 'sphere of vulnerability' of the primary patient meet and

overlap.There is thus a general pattern which underlies all the examples 1 will be dis­

cussing in rhe ensuing sections; all of which arc really no more than variationson the pattern shown in Fig. 3.8.Ifl. That is to say, we are concerned withrebtions between 'primary' agents and patients (artists, recipients) who figure,so to speak, at the points of origin and termination of art-mediated chains oftransactions. These transactions manifest themselves in the causal milieu whichthey botb share, in the form of 'secondary' or deriV:Hive agents and patientswhich arc indexes, objectifications of agcncy distributed in the causal milieu:

Primary agent --------; (secondary patient --t secondary agent) --------;primary patient.

However, the interest of all this does not lie simply, I hope, in the elaborationof abstract models such as these. Let us return to a more informal explorationof the cells shown in Table I, making use of relevant examples.

by thc artist and to be an index of his agcncy as an imaginer of the appearancesof things. William make's well-known pencil-sketch of The G/lIIsl lira Flea isa pure case of im:lginary :lrt, and a pure case of artistic agency dictating theappearance of a (fictional) cntity. Any picture of a unicorn is also 'imaginary'in that, as unicorns are as fictional as the ghosts of fleas, no artist has patientlyattempted to delineate their forms 'from life'. On the other hand, the unicornis a 'rcceived image' which was not invented by the agcncy of any known artistwho has represemed a unicorn. An artist who depicts a unicorn is not dictat­ing the form of the prototype, even though the prototype is fictional, so thisis not Artist-A -----7 Prototype-P, in the 'pure' sense. L\'leanwllile, each artistmay invent various details of the unicorn he represents, so, in those particu­lar respects, the formula would be satisfied. The inverse of Artist-A --------;Prototype-P is:

2. Prototype-A -----7 Artist-P, which is the formula for 'realist' image­making. Here, the appearance of the prototype dictates what the artist does. Iintroduced this concept of 'realistic depiction' earlier under the heading of'Prototype-A (Duke of\Vellington) --------; Index-P (his portrait by Goya). Theprototype, as social agent, in this case, impresses his/her/its appearance on theindex, via the mediating agency of the artist, who is a 'patient' with respectto the prototypc while remaining an 'agent' with respect to the index. I.n sum,the pair of expressions Artist-A --------; Prototype-P versus Prototype-A --------;Artist-P encodes the basic contrast between artistic activity as the originationof appearances versus artistic activity as the 'realistic' depiction or imitation of'given' 'lppearances. In practice, any given index may motivate the abductionof both of these formuhle; that is, in certain respects, the index shows theartist's imagination at work, causing the prototype to have a p,lrticular appear­ance, while in other respects, the index shows the prototype causing the artistto reproduce, passively, its 'given' appearance.

3. Artist-A ------t Recipient-P. This formula expresses the power of the artistas a social agent over the recipient as a social patient. Many works of art inspirewonder, awe, lear, and other powerful emotions in thc spectator. Artists,whose technic,ll prowess enables them to produce these powerful effects onrecipients, arc (sometimes) heroes, magicians, persons of power and consequence(see e.g. Kris and Kurtz ll)29; Forge 1966; :\'lorphy 1991). The particularnature of the awe aroused by :lrtistic activity is taken up below in a separateSection (5.2, helow), besides whieh I have already written at some length onthis topic elsewhere (Oell 1992h, 1993). The reciprocal of the artist-as-hcro(Artist-A ------t Recipienr-P) is:

4-. Recipient-A --------; Artist-P, the formula for the 'artist as artisan', that is,a hired hand who does the recipient's bidding. Here the rccipicnt figures as the'patron' (see above, ].6) rather than as the passive spectator. The pair of for­mulae Anist-A --------; Recipient-P/Recipient-A -----7 Artist-ll, as with the pairArtist-A --------; Prototype-P/ProlOtype-A -----7 Artist-P may hoth be abductedfrom the same index simultaneously. That is to say, from one point of view,

FIG. j.8. 1/ I. The indc~ as thepivot of th" an nexus

the Index

; the primary agent, the artist. as an intentional

tleing

p

A = the causal milieu over wt1ich the artist mayexercise agency. the sphere of causal control

and influence

"the primary patient/recipient (spectator) as anIntentional being

'" the causal milieu of the patient. Ihe sphere withinwhich the patient is vulnerable to control andinfluence via outside agents

A + P "the region whefe the sphere of agerocy of Aoverlaps the sphere of patiency (wlnerability) of:?The Index i$ located in this region,

40 ·I'h<, Art Nexlls ill/d!llt II/dcol" 41the index rna\' manifest the independent agency of the artist and his pre­eminence ave; the spectator, while the S:lme index, [rom another point of vicwmanifest... the subordination of the artiSt :md the pre-eminence of the patron.

5. Prototype-A ----t Recipient-P. One might call.this {h~ '.idol' .formula.Ilerc the agency abducted from the index, by the patICnl-rCClp~em, IS that ofthe prototype, who, besides causing the index to ass~me.a ccrt:un ap~r.ancc,

exercises social ;lCTcnc.... t:is-a-vu the recipient. A typu..':I.1 Instance of thlS IS thepractice of dicrat:)rs, 'such as M:l.o and Stalin, i~ ha\ing cno~ous imagl'S ~fthemselvcs posted on walls, keeping the population under conl1nuo~surHll­lance :lnd control \"ia their images. Later on in this book, J shall diSCUSS theworship of images in considenblc detail, so there is no need ro. adduce d~railedexamples at this srage. Ilowc,rer, it is important to nOle lhat thIS formul~ IS also'reversible' in the same way as the ones I deah wilh a mument ago. The IIl\-erscof ProLOtype-A -------t R/..'Cipient-P is:

6. Recipient-A -----t Prototype-Po This is the 'volt sorcer}~' formu~. ~oltsorccrv is the practice of inflicting harm on the prototype ofan mdex by mflIct­ing ha~m on the index; for example, sticking pins into a \\'3X image of the pro­totype. Volt sorcery will playa large part in the argumenr Ial,er.' as well, soI shall not give any more derails at this stage. III general ReCIpIent-A -------t

Prototype-P refers to sitU3liollS in which the protolype can be 'got at' in somewav \"i~ his or her im:lge, This may be as a result of malign artistic agency (aCls~ of Artist-A -------t Rccipient-P) or the agency may stem nOl from the anist'sactivit\" in making the imagc, bUl the recipient's activity in defacing if. Paintinga mou;tache on a picture of I\<I rs Tharcher is nOl necessarily 'arristic agency' somuch as a (hostile) mode of reception, through which the recipients of !\hsThatcher's image can obtain redress against the woman (prorotypc) rhey hatcand despise. This receprion-t.:lcric is nor necessarily mystical, that is, basedon thc supposition of 's~mpathetic magic' in the manner of \"oh sorcery. Aposter-<lefacer might quite f:llionally suppose that Mrs Thalcher herself, or atleast some of her supporlers, might see the defaccd poster a~d feel bad as aresull ofbein"" made aware of the extenr of anti-Thatcher feci mg.

It will be :pparcnt that I'rorol~-pe-A --------? Rccipiem-P{Recipient-A -------t

ProrOlype-P form a couple; a single image can be an index o~ bUlh of theserelations simullaneousl)'. Thus an idol is simultaneously an ~ndex .Ih~oughwhich Ihe god mediates his agency over his dewtces, who submIt' fO hIm In theform of his image; but at the same time, the devotees ,1elu~llIy have power overthe god via his image, bcc;\Use it is they who have m,ldc, msrane~, and con~e­crated the idol, it is they who offer sacrifices and praycI:s etc., wIthout' wh:~hthe god would hardly be so eonsequ~ntia1. In fa~t, there. IS, :~ gr~at ~~al mo.re, I,I~common between volt sorcery and Idol-worship t'han llllt1aJl~ meets the e)e

(see below, Ch. 7)· ...This completes our survey of the cells III Table I with the exception of the

cells showing 'sdr-redproc;ll' agency. To these I shall now lurn.

3. 10. /lldex-A ----? j,ulex-P

~n in~e.x I..m be seen as the 'cause' of itself. To c(II1\"e)' an idea of this, ima _me being a Spec~:Hor ~t a performance of the Chinese State Circus's acrobati;tcam. A~ :lcenam. POInt all the :lcrobats stan clambering II\-er one another andlo~-as If by magic they rul\·e turnl.'<! rhemseh·cs infO a majestic human pyra­mId. But who or \Vha.t has made this pyramid? C1e:l.rly, the acrobatics tC'.Im.And o~ whar ~oes thiS p~ramid con.sisl?-the acrobatics team. The human~~:ramld as .an mde~ (and a kind of work of art) is a 'patient' in the sense thalIt IS :-o~ethmg thai IS made hy someone (a collecti\·e someone, in Ihis instance)?ut It .IS .also an 'agent' in that the al."[ of 'making' is one that it performs onItself: I~ IS s~lf~lllade. There are many works ofart (inde.'\cs) which h:l\·e ehar­aetenstlCs ~lI111lar to a human pyramid crl.":lted by acrohats. For instance, longyaOls are dlspla~'e<l at annual festivals by the r\belam of the Sepik district NewGUIllC:t, as cult objects, They arc in fact decorafl,.'(1 (painted and provided withmasks) but the object on display is the yam itself, ralher than the mask. Yllmsgrow themsclve.s, It is true that yam-growers c:m :lssist yams to grow, t'echni<.:­ally, ,b~' hol1owmg out ~he earth around the growing tuber, and socially, byre.fral~mg from sexual Intercourse, which is deleterious (or more precisely,offenslve).to yams. The yam must be magically protected, but the magic of~am-gro~\rmgd~s not cause tuberous growth. The powers of growth inherenttn yam.s IS pre.clscly why they arc cultivaled cerellloniall) and exhibited; theyarc ob,ects of wonderment, attaining, somelimes, lengths of over ten feet.Yams of rhese dimensions are utterly inedible, their only destil1\' is to be lookedal and t~ be a. source of planting material (rams are, of course,-alivc and socialagenlS, Just like jX.'Ople). The ethnographer (Forge 1966) is quitc explicit inst~tin.g that ~am... are 'art objects' Cltcgonl.'all) assimilated to the sculptures andpamtmg whIch the Abelam also make and display.

Abelam yams pro\·ide a suitable example of indexes which exert agency withrespect to rhemsel\"(~. This is the abductive inference drawn by the Abelambut it is not in the lea... ! obscure; all living things are agents \\~ith respect t~themseh-es in rhat rheir growrh and form mav be attributed to their ownagency.. What is coume,r-inruiti'T, from our poin't of \'iew is that 'yams' shouldbe conSidered person-like agenls and 'works of art', But of course horticultur­a,lists frequently d~ personify their plams ;md the blooming back-gardens ofEngland abound WII h unacknowledged animist's. The behaviour ofattenders atg3rden sho\~rs .is exactly comparable 10 the behaviour or spectators at art showsexcepl. that H IS gCllel'Jllly less self-conscious ,me! solemn. Highly nuanccd ,Ies­thetlc judgemems ;Ire heel)' voiced on Ihe subjecl of roses and cauliflowers by~o-nonse_nse,malr~,~s I,:ho \\'.ould hardly <:arc \() utter any opinions al all 0;1works 01 art exphcnly ICknnfied as such. Such is the nature of our an world

which is no more ration31 rh:m the one oper..Hed b\" the Abelam, but which i~

predicated on a different SCi or social re1ationships~ocial class relationships in

42 Thr Art ,,",,,(XIIS alld /he blll"x Thl: Art Nexus and fill: JIIJU 43

is the subject-mattcr of the most developed branch of the cognitive psychologyof art, lhe part developed from Gestalt psychology by Rudolf Arnheim in hisclassic treal'ise on an and \'isual perception (J97~). Fig, 3,10/2 comes from theopening pages of Arnheim's work, and it demonstrates a fundamental visual­cogniti\'c phenomenon. Whereas if the black disc were positioned in the cen­n:c of the white square it would appear to be at rest, as it is in the figure thediSC secms to be drmm towards the right against some kind of resistance ortension (ibid, 12-13). Pans of indexes (such as the disc, though of course,Amhcim does nol usc this language) arc shown to he al the conductors of pic­torial 'forccs' affecting the appearance of balance encT"'~V <Trowth dvnamics, ",.'" '. ,etc. (passim). r have no intention of summarizing :\ I'l1heim's aurhoritative pre-sentation of visual psychology at this stage; the importam point to note is thatthe idea of agency intcl'l1al (Q the (pictorial or sculptuf:ll) index is an exceed­ingly familiar olle.. Abstract art exploits our perception of internal agency (or to be more pre­

CIse, cause and effecl) wilhin the index to a great degree and in fairly obvious\\ aJs. Patches of colour seem to whirl around, hover, dash, and fragment as ifthey had internal sources of energy and were engaged in complex caU'ia1 inter­actions. With representational al'l the situation is different, in that we ha\'eto distinguish between the 'internal' causal domain of the picture surfaceor sculptural form, and the extcrnal causal prOcx......"es in the world to which~he index relales. One of rhe moSI striking examples of 'apparent causality'III Western art are Ihe perrecll}' modelled depressions in the soft nesh ofPersephone's thigh, produced by the fingers of Pluto grasping her in Bernini'smasterpiece of illusionistic sculpture (see Fig. 3.10/3). These m.lrblc depres­sions are 'representations' of the callsal nexus be1ween gripping fingers amiyielding flesh; bUI I do not think we see them as such. Instead-so compellingis the illusion-we see these depressions as instances, rather than representa­[ions, of C'"J.usalil}. This kind of lrompe /'te,-t pseudo-causality (agcm/paticm

FIG. j. loll. The self_mati", indc.'f

parricular (l3ourdieu 1984). We find it hard to see things which 'gro..... by Ihem­seln~' as worl..s ofarl because. for us, the aeth-ities ofan 'arlisr' arc intrinsic to

the concepl ofart jrself. But from the standpoint of the anthropolQb'Y ofart thisis merelv a relative maner, an axis ofmmparison between different art systems.

1100\'~"er we do not ha\"c to consider onh' indexes which acrually 'make, .themselves' under this heading. Agency is not just 'making' but any modalitythrough which something affects something clse. Indexes do commonly­indeed universally perhaps-exert agency over themselves in the sense thatthey consist of visual parts, and these parts arc seen as :liTeering one anotheriorcmallv to the index.

To re~'er1 LO the human pyramid formed hy the acrobatics team; when wesee this form we recognize lhat it is in stable equilibrium. Each aerohal is exert­ing the neccssary force to maimain this equilibrium, bur ifone should make anunexpccted movement, then we should fear to see lhe whole pynmid mllapsc.The forces in the structure, lhe agl-llcy exercised by one part (one acrobat)with respect to the others, arc \ isuall~' embodied in the structure as a whole(Fig. 3.10/1). \Vhat \\e sec is a complex network of agent/patient relationshipsbetween indi\'idual acrobats, pairs and triads of acrobats, etc. wilhi" the index.Thus :lcrobat 6 can be seen as the agent who holds up 8 and 9, and rhe patientwhom 2 ;tnd 3 hold up. Together with 5 and 7, he is part of a three-man team,jointly holding up 8, 9, and ro, and jointly sustained by I, 2, 3, and ~, and soon. Innumerable relations of this kind can be extracted.

The important point is that it is not just in connection with works of anwhose 'parts' consist of hUl11nrl acrobats thnt such part-to-part 'l,nd Jlart-t~~

whole agent/patient rclarions can be extracted (or abducted). 1 he same IS

equally true of artefactual indexes ofall kinds. indeed, ~he ki.nd o~ pa.rt~to-pnrtand parHQ-wholc agcnlfparicnt (cause and dfeet) relatIOnships wllhm IIldexcs

• fIG. J.10/1. The black disc looks as ifil ·"·~rllcd' to relurn 10 the cenlre ofthe ~qllare. Sourcr: Arnheim 1974

3.1 I. ArfLI'I-A~ Artist-P

tions, whether they arc only 'suggest'ed'~.ls in patterns and abstr:lct art-orwhether they are depicted directly, as in Bernini's Pllliu and PerscpllllllC, testifyto complex intentional agency not in the index itself, but 'off-stage', in thecunning mind of the artist.

45The An /\,'exu_1 lind /he Index

Having explicated the sense in which an index, or candidate work of art, canexercise agency with respect to itself, let me turn to lhe kind of self-reciprocalagency which can he exercised by arrists. Every artist is a patient with respect'to the agency s/he exercises, indeed, artistic agency cannOl proceed otherwise.Consider lhe act of drawing something that one has not attempted to drawbefore (a Chippcndale chair, say). One desires to make an index which willrefer to this chair. The act of drawing is preceded (whether the ohject to bedrawn is present or not) by an act of visualization of the drawing to be made.One internally rehearses the line(s) which mUSl be produced, and then drawsthem (a drawing is always rcally a drawing of a drawing, thc drawing in one'shead). Because one's hand is not actually directly controlled by the visualizedor anticipated line that one wants to dmw, but by some mysterious muscularalchemy which is unerly opaque to introspection, the line which appears on thcpaper is always something ofa surprise. At' this point one is a spectator ofone'sown efTorts at drawing; that is, one has bccome a patient. Subliminally, oneasks, 'would I recognize [his (index) as the chair I wanted to draw?' just as if ithad been drawn by somebody else.

Drawing, and most other artistic skills (can-ing, etc,) arc what are known as'ballistic' activities, muscubr performances which t'ake place at a rate such thatcognitive processing of the 'outcome' of action only takes place after thc :lCtis complete, not while il is in progress. (The archetyplll 'ballistic' behaviour isthrowing.) Most often, if one is not very good at drawing, the result of one'sballistic chair-drawing gestures are frustrating: 'this chair is not the one Twished to draw-the legs arc too long and it is all lop-sided'. The patient posi­tion or the would-be artist who cannot draw objects 'as intended' is a fami1i:lrpredicament. Occasionally though, by a happy muscular fluke, the line drawnis actually superior to the one visualized beforehand,

This is the 'generate and test' sequence which is a fun&lmental feature of allcomplex cognitivc performanccs. Dennen quotes ValCry as saying 'It takes twoto invent anything. The one makes up the combinations; the other chooscs,recognises what he wishes and what is important to him in the mass of thingswhich the former has imparl'ed to him. \Vhat we call genius is much less thework of the first one than Ihe readiness of the second one to b'Tasp the value ofwhat has been laid before him and to choose it' (Dennetl 197'-): 71). Dcnnendevol'es a chapter to vindicating Valery's argument, I'hough he disagTccs that'choosers' arc necessarily more important lhan 'generatOTs'. Valery is ~bviously

riG, ,:;.' 013, C\u,aliry made,.isiblc: detail from PIn/IIand Persephi!ne b}' Bernini,SQnne: H. Hibbard (11)76),Bemi"i (2ml cdn.;llarmondsworth: j'enguin),plate ,6

The Arl Nexus lind die Illdex

interaction) within the index is not necessarily confine.d to illu~ioni.sticWeslernart. Any representational index, which depicts cal.lsal mteract~ons m th.e prot~­type, can also be considered as a separate domam of causal,Ity u.nto Itself, mwhich parts of the index causally interact with other parts of the 11ldex.

However, in the light of our prcvious discussion of primary and secondaryagency, it is necessary to qualify the above in certain respects., Abstract pat­terns ~lppear to show 'cause and effeet' rclations between monfs ~ather than'agent/patient' relations between motifs, in that nothing ma~es us.thmk ~h;lt .t~emotifs in patterns are sentient in themseh·es, rh:'t they have mtcnllo~s 01 ~~sl.resetc. \Vhereas the part-TO-whole relations withm the human pyramId test~fy to

'pnmarv' intentionality on the part of acrobats wJlO.m we instantly re~og?l.ze asintentio"nal agents, the same is not lrue of the relatIOns hem·.een the lIldl\'ld~,dflowCT motifs on our floral drapes, be they e\'er so bustlmg. !hese. mot.lfsonlv have 'seconc!arv' agencv, they manifest the effect of agenCY!llltentl?nahtywirilOut possessing i't intrinsically. These motifs only interact causa~l'y \Vlt~ oneanother, not intentionally. Howe\'er, even il~ thi~ case, we do see Il1tentlOl~alactivitv' hcre, but it is displaced onto the lluaglllary ereator of the pattel n,rathcr"than onto lhc physical constintents of the pattern. Complex causal re1a-

TIt( Arl Na'us (/ml/ht ltll/l'X

talking about Artist-A~ l\rtist-P in precisely our sense. Of course, he isspeaking of poetry, an art form which pennits indefinitely many cycles of sclf­correction. This is not always possible in graphic/plastic art, where erasure bythe 'chooser' may be difficult or impossible, and the whole enterprise may ha\'eto be resrarted if the resuk" of any given 'test' cycle are deemed uns:uisfact­Dry. However, one docs see copious e\'idence of 'generJ.te and test' in artists'preparatory drowings, especially those by such Renaissance masters of the artof drawing as Raphael and Michelangelo, many of whose exquisit.cly drawnforms emerge out of clouds of provisional strokes known as abbozzi.

Moreover, it is often the case, especially witll more complicated drawings,paintings, or carvings thar the fin.ll product comes as a surprise to the artistsimply because it never was rhe 'final product' which was visualized before­hand, but only the successive generatc-and-t~l cycles along the road to itscompletion. D'Aze\'edo cites the testimony of an African {.":In·er, who says:

I sec the thing I ha\'e made [a Sand~ mask\ coming 0\11 of the \lomen's hush. It is nowa proud manjlr/(J lspirit] wilh plent~· of women runninl; after him. It is nOI possible 10see anything more wonderful in this world. Hi.s f:l.ce is shining, he Iook.s this \la}' amIthat, and all th~ people \londer aboullhis beauriful and lerrible thing. To me, il is likewhal I see when I am dreaming. I S:t~ 10 mlself, this is I\'hat my ntmr rfamiliar spirilJhas brought into m~' mind. I say, I have madc this. I-Io\\' (:an a man make sut:h a thing?It is a fcarfullhing that I can do. :-.10 other man can do it unles.." he has the right lno\\"­ledge. No woman can do iL I feel that I have borne children. (d'AzC\'edo in Forge

t97J: 148)

The artist v;lcillates between the 'patient' response, the astonishment and awethat the Sande mask product's in him-'I-Iow can a man make such a thing?'­and self-approbation stemming from Ihe fact that, after all, his was the agencywhich produced it-'It is a fearful thing I can do.' II would be impossible to

find a morc explicit instance of self-reciprocal artistic agency than this. Thecarver's statement that in caf\'ing the mask, he thinks he has 'home children'and his e\"idenu\' total commitment to the notion that the mask is a living,perceiving, being 'looking this wa~' and that' is also very helpful testimonybolstering our general hypolhesis that, anthropologil.:ally spc3king, works ofart are best considered as t~ pes of agents. The makers of our idols are no lessenthusiastic idolaters than the rest of us, bel..":luse, in some sense, they 'llwaysrcmain passive spectators at the birth of their \'cry own creations. Correctlyexprcssed, this is really rfArrist-AJ ~ Index-Al ~.Artis.t-P.. .

Finallv on this subject, I should signal a theme whIch mil not fully surfacefor a lo~~ while yet. Artists do not just produce singular 'works', thcy havecareers and prod~ce an «livre (I will reservc Ihe French word for 'work: tomean 'lifetime work' or 'all the work to dale' of an arrist). Artists arc not lustpatients \I-ilh resp{.'C1 to the '\\ork' Ihey are producing right no\\·. They mayalso be in the patient position v;s-a~vis all the work they ha\'e e\"cr produced.

17,r Art Nr.\"III (lml lilt 11/110' 47

Some anisrs, for instance, seem unwilling 10 repeal themselvcs, or work forpatrons who d.emand that they should not do 101). For instance, it seems tome that POUSSIn, for ~\'hate\'cr reason, rarely if e\'cr repeatcd a compositiona~ all closely. If.PouS:-ln, as I hazard, obscn'ed a principle which discouraged~~~ .from rc~.tlng hIS own work, then his (;\"ery work, indh'idualJy, was negat­I\d} determmcd by.all rhe others, so as not to resemble them in composition(though of course hiS personal style remained rel.uh·c1}' consistent). Renoiron. the oth.er ha~d (as flne may \'erify by \-isiting [he Barnes collection, i~Phlladelphla~, painted a large number of 'bathers' which resemblc onc another\'ery closely Indeed; and ~c di.d so, presumably, beC,luse he had (in Mr Barnes)a patron who was happIest If The next paiming he purchased frOIll Renoirr~embled .alllhe ones he had purchased frolll the same artist on previous OCca­sIOns. In eIther ca.<;e, the anisl was in the 'patient' position with respeCI 10 histotal (£Ilvre at any gi\'en time, to the extent that his Current work had to berelated, in a specific way, to his pl'evious works.

3· 12. Recipient-A~ Reo'pient-P

The category of 'recipients' splits into agents and paLients in a \·er,· salientfashion, so much so that one might be tempted to den\' that it was'realh asingle l..":ltegory at all. The differentiation that I h,1\'e in ~lind is that belw~en'passive spectators' (the general aTt public) and 'patl'Olls'-thosc who actuallycommission a~tists to ~roduce works of art, and whose agency, as patrons, i~~consequently IIldexed III the works of art they have caused to come into exis­tence. Patronage is a very signific:lm form of agency from the point of vic\\' of~he anthro~lo~yofart. I.t Sl..'ems very diHerent from mere spectatorship, whichmmlvcs hemg In the patient position vis-ii-vis a work ofarl and being 'clused'to respond by it (being impressed, fascinated, etc.). On the olher hand, artpatrons are profoundly impressed, or can be, by the works ofan that thc\' have~uscd to corn~ into existence by commissioning them. The Sande ;depts(lmponant, SClllor, women) who commissioned the caner whose words arecir~-d a~\:e, arc the same women \\ hom he dCS(.Tibcs as 'running after' it, subjectto Its splTltual and masculine allure. Here, for comparison, is a quotation froma fourteenth-<:enrury observer, describing the public homage paid to Dueeio's,Hajes~yofClirist altarpiece for Siena cathedral by the patrons of the work:

.~nd on thc day. that it [the nell' pai11ling] was carried to Ihc Duomo the shops wereshut, and the 1~lshop conducted it gn:at and devour company of pricsts and friars insolemn proceSSIOn, a(;companied by lhe nine signiors, and all lhe Ilffi(;crs of the eOlll­nH1~e, and aillhe people, and one aftcr ;\Il0lher the lI'orlhil:.\t wiTh lighted candles inIhclr hands look pla<."Cs ncar the picture, and behind came the women and children II ilh~'TCal devOIiol1. And they an:omp:lnied the said pinurc up 10 the Duomo makin" lheproc"eSsion around the ('..ampo, as is Ihe custom, all the bells rimcing jo\~ush" O~I ofrt:"ercncc for so noble a picture as is this. (J\ IS ofCo 1311, cited in Holt \"~1. i 11)57: 135)

nh' Arl Ntxus lind Iht {/lila The Af/ No,m (wd till! {lIde.I' 49The bishop and lay iluthorities (the 'nine signiors') who commiss.ioned t~ework, glorified themselvcs by conspicuously showing reverence, m pubhe,towards the product of their own agency (mediated by Duct:io). They were, ofcourse, also revering Christ, the Virgin, and lhe Saints, but it wa~ really thepicture itself which was the object towards which re\'t,~rence was bemg shm~·~,

because, at that time, such holy icons were belie\'ed to ~rotcc.t the specificinterests of the commune which harboured them. DucclO's plclure was toreplace an older icon, the Mallo,ma of Iht LA'?' EJ'ts (remo~·ed to the churchof St Bonifacio), to \, hich was auributed the victor) of the Slcncse forces O\·erthe Florentines al the recent banlc of iHonte Aperto. So il \\as most importantlhat the new picture should 'realize' thaI the whole town was depending on itto 'keep us from the hands of traitors and enemies of Sicn;t' (ibid.).

In olher words, the yery essence of successful performance of the 'palron'role, necessitates a show of reverence wwards the products of patronage. Th.cpatron is primus ;I/Ier pares ,lmong the general art public. L;nless the patron ISvisibly and/or privately il11pres"ed by the index of which slhe is p.atron, th~ veryact of patronage is a failure, and the resources which h:lVe been 1.nvc.ste? ltl thecommission 11:lve been w:lsled. It follows that patronage has, mtrtnslCally, aphase in which the palron/agent rRecipient-A] is a p:lticnt [Reeipient-P].

3,13. Prototype-A~ Prototype-P

The prototype ofan index can be a patient with respect to the i~de:" which, byrepresenting him or her, incorporales his or her agency: Con~lder th~ case ofCouncillor (later Mayor) H, hailing from an old-est'abhsh(:d mdustnal town,Councillor H's party is more or less permanently in power in the town, and1-1 \\ ho is an intelligent and efficient busincssman, rises gradually throughm~mbcrship and chairmanship of committees to the position of leader of theCouncil, which he occupies successfully for a number of yC;lrs. The town flour­ishcs· at Ihe appropriale time, he relinquishes the leadership 10 a younger(;01I~gue, and accepts the mayorahy. lie presides \\ith dignity m:er publicfunctions, and, behind the scenes, helps to secure no end of lucra~Ivc.grantSand contracts for his eommuniry. He is uni,'ersally popular, even wlIlnlllg therespect of his sometime political opponents. To commcn~or~le h.is term ~smayor, the Council un:ll1imously propose th:lt he should sit for hiS portraIt,which will hang in the council chamber, in an honoured place. He agrees,partly because he knows how pleased his wife will be, ;1I1.d ~)anl~' bec,aus~ .~:obtains an assurance that he will not have to endure long slttmgs 111 t~e artist sstudio which as a srill very husy man, he has no time for. And so It proves;~he ar;ist onl\: needs an hour to dash off a number of quick sketches, and takea large num~r of photograph" from various angles and distances.

The appointed day arrives; the Council is assembled, ;llld, sea.te(~ on tl~emayoral throne, H w;ltehcs as the curtain is drawn :lparl and hiS Image IS

revcaled. :!\s the in~'·i.table., and prolonged, speeches proceed, H has a goodopportunity to SCrutllllze hIS portrail. As he docs so, he is assailed b,- panic anddespair. For the portrait appears to him not to be a reprcsematio~ of a manbUl of some vegetable, a turnip in fact, with a revohing purplish sheen to it'disfigured by nameless appendagl..'S, He is not a vain man; in fact it is rea~since he looked at himself for more than a moment at a time, for he is ·e\·enaCClL<.tomed to sh'l\·c in the back of his limousine on the wa\' to work ratherthan waste precious momenlS before the bathroom mirror. But it is a ;erribleshock to him to discO\"er how ugly he really is, and the effect is made worse b}the contrast between his impossible turnip-head and the splendour ofthe may­oral costume he wears, He cannot blame the portraitist for his discomfiturehe(:ause he is intellib'"CllI enough [Q appreciate that the portrait is faithful to hi~actual appe'.lranee; the anist did not wield either his camera or his brushesinexpertly. Would that he had!---but II, who bebran life as a~ apprentice andimbibed the Protestant ethic, knows honest workmanship when he sees if. 'Isthis how I am goingro be remembered forever,' he wonders, 'as a turnip? Whatdoes my appearance maHer anyway? Why docs my jace, of all the things !O dowith me, have to be remembered?' He wishes he could have been memorial­ized aniconicillly, by something in the nature of an inscribed plaque-but youhave to be dead 1'01' 'hat. In the cnd, there is nobody he can blame for what hashappened but himself, his own ugly mug. Ifhe wer~ bener-Iooking, his portraitwould not have been such a horrid lhing, As il is, II is his own victim, lhc vic­tim of the direct causal influence his actual appearance [Pl1)totype-A] has of theactual appearance of his porrr.lit, which is so dam:lging to him [prolOtypc-P].

This, I admit, is an invented example. But there arc numerOlL" real instancesof sillen; for portraits feeling victimized. I might cife the wdl-documenlcdamipathy felt by Winston Churchill towards the portrait made of him byGraham Sutherland (whose public circulation he prc\·emed). This ponnit iswidely regarded b) critics as a penetrating study of the great leader, "cry 'true10 life'. Churchill himself infinitely preferred the heroic photographic portrailby Karsh, as well he mighl han, in dcfiance of contemporary opinion whichdetects much les" aUlhenticity in Karsh's photograph than in Sutherland'spainting. Churchill was vain cnough ttl blame Sutherland, in public, for theuglincs.'i of his image, ralher than himself, unlike nur honest H, who knowswhere agency rc<lll)' lies in the coming-in to-being of ugly portraits. But I thinkhe must have had private doubts-otherwise his reaction would not ha\'e beenso violent.

This type of self-reciprocal agencY/paliency exercised by the prototype ofan index with regard to itself is act wIlly very r,ulliliar to us. If we look intothe mirror, and dislike whaT we sec, or indeed approve of what we sec, we arcresponding, ;IS patients, 10 an index (the mirmr image) of II hich we are theagents, Portraitme is only a "recial instance of this, mediated by the :Ictil"iliesof an artist, or;1 pholOgr.lpher. Wherel"er there is really or supposedl) a <--:tus:ll

4·1. Hierarchical Embeddi/lg OJAgettl/pfllie1l1 Relations

This completcs our survey ofbin,lry relations. As the reader will have noticed,in the discussion of these rebtiolls, it has often been necessary [Q refer to morecomplil;ated relations, involving more than two terms. An example of such anexpression, would be:

[[ [Prototype-A} -7Artist-Aj-7lndex-A]~ Recipient-P.

This expression refers to a nexus of agent/patient relationships such that therecipient is the 'patient' and the agent acting on him is the index. This is therelationship between a (secondary) agent (the index) and a 'primary' patient, inthis instance, the recipient. I adopt the graphic convention of alwa)s indicat­ing the relationship between the index-agent and the 'primary' patient in arelation by the usc of a long arrow '----t' as opposed to the short arrow '-7'indicating subordinate agel1tfpaticnt relations. &:C:t.llse of the centrality of theindex (sec aoo\'e, 3.8) it is always immediately to the left (or occasionally 10

The right) of the long arrow. Agents arc always placed TO the left of patients;the terminations '-A' and '-P' are really redundant because any term 10 theleft of another is always interpreted as an 'agent' with respect to it; however, Iretain the '-A' and '-P' suffi.'\es because they make the resulting formulae morereadily intelligible, or at least, I hope they do.

The index in (he above fonnula is not acting on the recipient autonomously.It may be the focal carrier ofagency, bm i( is serving to mediate other lypes ofagency affecting the patientJrecipient. The recipient's response to the indexincorporatcs the abduction that the index is a 'made thing', the outcome of theagency of an artist. That is to say, (he index is an agent with respect to therecipient by \'irrue of the fact that the recipient abducts the agency of thc arlistfrom it. The index is an agent (with respect 10 the recipient) but it is simulta­neously a patient, with respel;t to the agency of the arrist, which it mediates.This 'indirect' relationship between the recipient as patietH and the artist asagent is expressed in our formula via lhe brackets. The term 'index' iI/eludeswilhill ilself.mother term, 'artist'; thus, '[Index]' expands to become 'l" [Anist]Index]'. Adding '-A' and '-1" suffixes, and the agency arrow indicating thatthe artist is :m agent with respect 10 (he index, this becomes: [ [Artist-A] --+Index-i\l----t Rccipient-P.

50 Thr Arl Nrxu$ and Iht' [,Itlrx

relationship such that the prototype is the cause of.thc index, it. must be that theindex is at least potentially able 10 cause effects (dismay, etc.) In ~he pr~totype.

A very pure 'artist-less' casc of Prototype-A ---: Protot)'pc-~ IS p~ovld~ bythe myth of I an:issus, whose beguilement by hIS own reflectIOn (mdex) In apool ~used him to fall in, and drown. However. m?Stly the eff~ts of the i?dexon the prototype are not primarily caused by the IOd~x~ bUl sImply med13t~dbv the index, and agency lies with the artist or the recipIent. In the Churchill'S-utherland case, Churchill considered himself the victim of the agency of theartist, rather than his own agency as an ugly pen>on.

4

The Involution of the Index In the Art exus

52 Tilt: IlIi,,,fulioll of/ht: Il/iian,t IRW/II/if/II oJ111t IlIdr.1' 53

Finally, in the aoo\'e formula, the promtype also makes an appearance as anagent with respect 10 the artist, the index, and the recipient. This can unlylx:eur when the abduction is madc that the activities of the artist arc subordin­ated to the prototype, for example, to the appearance of the prototype, as inrealistic forms of art, such as ponr.liwrc, From a certain point of\'iew, a por­trait is an index of the appearancc of the siner, mediated by the artist's per­formam;c in creating an index, which mediates the prototype's appcaram:c to therecipient. The sitler's appearance c~lUsed the index to appear as it does. This isexpressed by enclosing the prototype in the 'artist' brackets, 'llArtisl] Index]'becomes '[[fPrototypeJ Artist] Indexf, which finally witb the addition of the'-A' and '-1" suffixes, and the agency arrows, becomes HlPrototype-AJ ~Artist-A] --) Index-A] --t Recipient-P, our starling-point.

4-.2. The Effici ofSubstilulions

\Vhat purpose docs such a formula serve? I am only tuo well aWare rhat manypeople find formalizarion objectionable, especially the kind of people whointerest themselves in artistic matters, many of whom (like me, in fact) sufTeredexceedingly during mathS lessons at school. I dare say all these symbols, eventhough I have kept them to the minimum and made them as perspicuous aspossible, seem to have little to do with 'art'. None the less, there is some pointin formalization if it genuinely assisls one in thinking clearly. The formulaunder discussion, I claim, encapsulates in the most L'(.·tmomical way,just 01lt ofthe myriad possibilities which exist for an objects to mediate social relations.

In more impressionistic languab'"C,

rrrprototypc-A] --t Artist-AJ....,. Index-A]~ Recipienr-P

pieks out the situation in which a 'p:lssive' spectator is causally .lffecled by theappearance (or other attributes) of a prototype of an artwork (the index), whenthis allribule is seen as itself causal of the specralOr's response. This is a verycommon situation. A good example of this would be our response to Reynolds'sportrait of Dr Johnson (Fig..p./I). This portrait, exccllent example lhough itis of Reynolds's art, is none the less seen primarily as an icon of Dr Johnson, aculture hero of the English. Re) nolds, in painting his portrail, is understoodby us to be as much in awe of the lexicographer as \\e are outSehes, and thishas affected the \\,:IY in which the sitter is portrayed. The situation is quiteotherwise in the case of a portrait, or ostcnsible portrait, such as lhe Mona Lisa,by Leonardo da Vinci. The priorities arc reversed in this instance; thc features,or some semblance of the features, once possessed by the woman rcferred to inLeon:lrdo's picture, are signific-.Illt only in so far :IS they mediate our awarenessLcon:lrdo's art as a painler:

[[ [Artist-A] ~ Prototype-A1~ Indcx-A1--t Recipient-P;

FIG..pll. The prolOI) pc asagent: S""'MdJlliinson h)·Rcyonlds. SlIurrr. TalCGallery, London

that is, Lc~nardo. is ~L'"Cn as responsible for the .\1ona Lisa's appearance, or at

lea~t,.wh~t IS ~1SCmatl?g and compelling :lbout her appearance from the patiL1ltlreClplen~ s pomt of VICW; whereas Reynolds is lIl// seen ~IS responsible for thecompcllmg aspects of Dr Johnson's appcarance.

By m~king s.ll~stitlltions within the formula-in this case, by switching[he rclatl\·e pOSIllon~ of the artist d th . 'h'"an e prototypc In an ot crwlse ldenncalformul~-we can express the basic differcnce between representations in whichthe ultm~a{e sourcc ofagency O\·cr the index is attributed to the artist (as in Ihe1\1ona LI~ case), and those representations in which ultimate a""enc\ seemsto ~est wIlh .Ihe prolOtype (as in the case of Dr Johnson). Our form~lae aredeSigned therefore to pro\'ide models which can be manipulated and trans-formed at will so as to disc·· b' II' ,. . ' . nmmate etween {/ pOSSIble ((JlllbmatiollS of agent!patient rehltlOllS between terms.

4.3. Tree-Slmc!lIres

Unde~lying our formulae are tree-structures as represcnled in Fig I Th"<YT h. .. I· . -1·3 I. IS",.ap ~c cOll\cntton IS ess economlL":lI, bUI more perspicuous than the for _lac usmg brackets. In parricular, it brings om the crucial idea that the 'in~~x

Tilt IIwol"litJl/ Qfthe II/de~' 77,( 1t/t)V/llliQI/ ofr//(' llItlex 55

[[Recipient-A] -----7 Index-A]---+ [Artist-P -----7 fPrOlOl)'pc-P]].

FiG. ~'312. The hicnrehie;llembeddcdncss of

,tgent-patient relations

~Agenl __ I':aUCflI

~Agenl • f'.tlient

~1St.'l:undaryl~ _ PatJ€nt

(tertla,y) Agent -r:llient

(prima!))

{primary)

~(prima!)) Agent • Patienl

~l~l "&ent _ P.lliL'fIl

art is, or;1.I letst wa,,,, indicati\'e of the lives children lead as 'patient.<;'. Schoolarl is produced at the behest of adults, so as 10 plcase them, or al least nUloffend them. Anthropologically spcakjng, the imporranl feature ofschool art iswhat it tells us about the social relations between adults in authoritv and thechildren in their charge. On the OIher hand, the teacher has not, on ~his occa­sion, told the children in the class what CO represent, so although c;lCh childauempts 10 paint something which will be acceptable to the reacher (no 'rude'people, no bleeding corpses, plenl)' of \Ciews of mountains and boraniealh' ques­tionable OO\\'cr-pieces), each is obliged to exercise agency within the 'patient'role. lIenee the requirements of the formula given above would be met.

Finally, let LiS take another instance, this time of !\rrist-!\~ [1ndex-P-----7[prototype-P -----71 Recipient-Pili. This formula shuws the anist as sale agent,exercising agem:y over the index, which mediates his agency over the prototype,which in tum mediates his agency ovcr the recipient. This, so to speak i<; thc'artistic genius' formula. A suitable example mighl be provided by the \:'ork ofSah-ador Dali, a paimer who played the 'genius' role 10 the hilt, b~fore an ador­ing public. An~ painting by Dali (e.g. Tile PersiSftllct o["Jelllo,.]' a.k.a. TIl( Soft

FIG.•.3/" The multiple le,'dsof agencl' within the index

Inde_~ ~........... Agenl • Palient

~Agent __ Patient

r""-\ARenl - PalLenl

/'base man'

as agent' encompas....cs, within itSelf, hierarchically subordinate 'patient' rela­tionships, and conyerscly, the index as patient contains subordiruued agencyrelaTions. The index, in OThcr words, has an in\'olute hierarchical structure,permitting abductions of agency at multiple levels simultaneously.

The uee-structure shown in Fig. 4.3/( is nm thc only possiblc one. For for­mulae with fOUf terms, such as the one we ha,'c been considering, there ate alsofour morc possibilities (Fig. +3/2).

An example of the second type of'trce'-lhe kind in which both agents andpatienrs arc to be found on either side of the 'primary' agent/patient relationswould be:

This corrcsponds to a situation in which the artist is a 'patient' with respect tothe index, which mediates the 'patient' relationship he has with the recipient;with respect to Ihe reference ofthe image, howc\'er, he is the agent. 'Recipicnt­A' in a formula like this, means generally the rt:cipient as patron or primemm-er. 'Artist-I" implies that the artist's passive acceptance of the pat'ron'sdemands on him are what we abduct from the index; 011 the other hand, theprocotype of the index (in this formula) is contributcd by the agency of lheartist. What kind of real-life index might motivale the abduction of agency'distributed' in this way? Well, consider a school situation as follows; rheteacher (recipient/patrot~) enters, and says: 'coday, class, I want you :111 ro paintsomething from your own imaginations, so gct busy! ... '. The young artistsaccordingly set to and produce their indexes-under orders. The resultingworks ofart index the agency of the teacher; but for Ihe teacher giving Ihe classit." inslructions, none of these exercise." in imaginali\'e art would exist. School

lnd~~...........~

Axent • Patient

~Agent _ f>.uit:.... t

r""-\AKent -P~tleT1l

/·~""'n·

The IIIi.:ol/lf;on of/he IlIdo; '11J~ bn"'OllIlio" oft/Ie I"dtx 57

Watch) can be seen, and frequently is SLoen, primarily as an index of his agencyas a painter. Dali's painterly agency is emphasized b~ the brilliant u.'ehnicalfinish of his work. !vloreonr, the prototype ofa Dali (the representational con­tent) is supposedly O:\li's pri\'ate drc-,lnt-world, f'Jther than anything in the rea~

extern:ll, world, which he was suiving 10 represent. Hence rhe prototype is inthe 'patient' position, relative to Dali's agency as the producer of I he index. TIlePersis/mee I)fAlclIIl)~)1 is a Surrealist 'self-portrait' of Dali. Finally, lhe recipientis not a patron, bill a passive spectator, whom D:lli regarded with a fine showof arislOcratic contempt. The aim of Dali's art was to dominate the spectatorby subverting and deranging his or her petit bourg(.'Ois sensibilil'ies. Sadistic,domineering artists like Dali do not target their an at a public offellow-sadisrs,but at a masochistic public which revel.. in being ourraged and loves theOpprt.'s.<;or. I-knce it is appropriate 10 indicate the recipient as a 'patient' here.

There are, according to my laborious calculations, thirty-si.x formulae whichcan be derived by combining index, artist, recipient, and prototype in agent!patiem rdations corresponding to the tree-structures shown in Figs.4·3/t-2,and keeping to the stipulation that the 'index' has TO figure either as the prim­ary agent or as the primary patient. Ilowever, the reader will be relieved to betold (hat I do not intend 10 provide instances of e"ery single one of them,rhough perhaps examples might be found. There would be linle point in doingso though; these formulae just provide a means of distinguishing between dif­ferent distributions of agent/patient relations in the \-icinity of works of art;they do not prediCl them or explain them. And, as we will sec, in order 10 pro­vide appropriatc models for very common modes of artistic agency, we need toadd some furlher refinements. There is, in particular, no empirical reason whyany of our four basic 'rerms' should appear only once in a gi\'en formula. Wealready know this, bec:luse in the inst:lnces of self-reciprocal agency discussedearlier, the same term necessarily OCl,;urS twice, as an agent 01' a patient. Butbefore I conclude the discussion of four-term formulae and their associatedtree-structures, there are some fUrlher points whieh require discussion.

The first of these is melhodological; it concerns the degree lO whieh eachfonnula is to be understood as a schematic dC'iCriprion of an 'objectively' dif­ferent situation, as opposed to a different 'perspective' on a situation whichremains the same. We can conveniently discuss this problem with referenceto lWO examples that have already b<.'Cn introduced; Dali's Soft WalCh and lheJ"IOIIII Lisll, by Leonardo, for which I g:l\'e the formula:

(I.eonardo) I [(Artist-A] ~ Pl'ototype-A] ~ Index-A]~ Recipient-P

while fol' the Dali I gave rhe formula:(Dali) Artist-A~ [Index-P --? [ProIOlYpe-P --? [Rccipielll-P]]].

The diITerence berwcen these two formulae is produced by a shift of the 'index'term from the left oflhe focal agency arrow (~) in the AI01I1l Usa fonnula,to the right, in the 50ft W(/uk formula. It i.s far from being my contention that

this shift results from any objeeti\'e feature present in either of these paintings.The shift is a shift in 'perspective' on the nexus of relationships around thesepaintings. The 'Dali' fonnula puts the spotlight on the artist; his person andacth'it)" arc out in lhe open and thematic from the obscn'er's poinl of \'iew. Allother factors in the siruarion are seen as suhordinated to him; thc canvase.... 011

show are, above all, 'Dalis'-passive impressions of his dominant personality,mediating his agency over his public. The cult of personality is, anthropologicallyspeaking, the salient social rransaction fl'Om this perspective. \-\ie could, how"­ever, refuse 10 take this pOlm of view; focusing instead on DaH's painting as theoven agent, rnther rhan on Dali the painter as the focus ofour attention. Thenour pcrspective \\ould correspond [Q the Malia Lisa fomm!a. Tlu PerrislenuofMemolJ' is a distinguished painting, desening of serious art-historil.."31 con­sideration independently ofthe cult ofpcrson:llity surrounding Dali. Con\'ersely,there is reason to suspect that Leonardo exerast.'d agency not only OVCT his paint~dt:anvases, but also in initialing a cuh of pcrsonaliry of his own, which is parr ofhis historical legacy, just as in the Dali case. From the point of view of the daVinci cultists, the Mona Lisa is thematically perceived, not as an image, but as asacred relic ofLeonardo, the semi-divine creative hero. So we may be at libertyto redescribe tl1c ncxus of relations surrounding the .r\1mUl Lisa from this altern­ative perspective, in terms of the fonnub we previously used for Dali.

What changes, and whal remains the same, if we make these substitutionsor redescriptions? In a sense, the difference bctwl.."CJl placing the index in theagent or patient position is rhetorical; akin, in fact, to thc rhetorical differencebetween (i) 'Leonardo da Vinci painled the Mona Lisa' and the equivalent 'pas­si\-c' construction (ii) 'The ,Halla Usa was painted by Leonardo da Vinci'. Al­though rt:sponsibility for the A'lolla Lisa is in both cases attributed to Leonardo,only in (i) is r,eonardo the subject or topic of the sentence, whereas in (ii) theMonti Lim is. Senrenee (i) corresponds to the formula in which the indexcomcs to the right of rhe 'agency' arrow, whereas semcnee (ii) corresponds tolhe L"Use where agency can be attrihuted to the index, :IS in: 'The MOlla Lisa(which was painted by Leonardo da Vinci), blew my mind an"Uy when I was akid'. In fact, passive constructions largely exist for use in relative clauses likethis. This sentence does not mean quire the same as one in which Leonardo daVinci figures as main subject: 'Leonardo da Vinci (who painted rhe MOllO Lisa)blew my mind away when I was a kid'. The same basic information is there,but a different syntacrical pattern implies a distinctly different 'analysis' of theworld. Which analysis is the appropriate one is a matter of social or psycho....logical judgement. On this basis I would argue thaI although the decision 10lreat the index as primary agent or primary patiem is a matter of choice notdictated by the 'basic fact's' of a situation, the choice is not arbitrary, but ismoti\'ared by sociological or psychological considerations of appropriateness.

The next poinr is related to this. In the formulae I ha\-e prc.<;ented, it hasbeen stipubted lhat lhe index is ah\a)s focal or central agent or patient.

Tht Inrolll/itlll f)flht Indo:

'Central' means 'visible', 'overt', 'immediate'; it docs not nect.'Ssarily mt.'an'most important'. In particular, it docs nOT mean, 'most endowed with originalagency' when the index is the agent, or 'ultimate object of agency' when it isthe p:lrient. In;l formula of the 'Leonardo' type:

UlArtist-AJ ~ ProWt) pe-A1~ Index-j\J -7 Rccipiem-P

Leonardo, as artiST, is well to the left of the arrow of primar) agenc), which, inthis case, is the agency of the index flyer the spectator, whose mind is beingaffected, thereby causing him later to remark 'The MOlin USII, which was paintedby I.etln:lrdo da Vinci. blew my mind ,lway when T was a kid'. I,ennardo, viahis origination oflbe appearance ofThe lady in his picture, and his representa~

tion of this appearance in the index, is Ihe 'hidden agent' to whom acc<......" isgained :It two removt.'S, primarily "ia the index, ;\I1d sccondaril} via the appt:ar­ancl': of the tidy represented in the index. r,eonardo is like God, invis­ible himself, bur visible via his \,·orks. i\nthropologisls will be very Elmiliarwith this kind oflhing. As a i\leI:lI1esianist,] find the Melanesian idiom of the'base' or 'root' (of something) springing to mind here. This idiom is used h)Melanesians to indicate the ultimate cause or origination of something, At apig-festival, for instance, where hundreds of people congreg<lte to I.:onduct pig­exchanges with one another, e,lch seemingly motivated by their own interestsand exchange partnerships, one or two leading men (such as Onka, a famous'big man' nfthe Mount Ilagen tribes in A. Strathcrn 1971) will be singled outas the base-men or ruot-men of the occasion, those \\hosc primar)' will andagency is manifcsted, not just in their own acts as pig-<:xchangcrs, but in theacts of will and agenq manifested by e\'erybody else present as well.

In the formula given above, I ,eonardo da Vinci is the b,lSe-I11Ml or root-man,since e\'en if the primary agent is the index ([rom which his agency is abductedat two remm'c...) he stands at the 'origin'. Schematically, therefore, in a brack­eled expn.:s...ion such :is:

rr[AJ -; Bj -; Cl

'A' is in 'base' position, or, we might even say, 'bass' position, since like thebas." line in music, the onc in Ihis position exercises mediated agency o\'er allascending le\·e1s. !-I1l\\e\·er, the analog} is perhaps inexact, in thai the sametype of tree-structure can be made to apply in Ihe case of patients :IS well asagents. Thus in the Dali formula, the recipients, Dali's public, suner hisagency through a double mediation, consisting, tirst of :til, of his dream-worldimagery, which disrupts their normal sense of the real, and secondly of histechnical mastery, \\ hich disrupL" their preconceptions about (Dali's) agency,sincc it seems supernalUral1~ proficient (sec Gel! 1992h). This gins risc to thein\·cr.;e structure:

Tht b,.vollll;on ofIk~ Jnt!~x 59

\\'hcr~ 'F' are t.he recipients in 'base' position as patients, and 'D' is the index:IS pnm,lrr patient.

There ~s nOlhing to preyent mulriple-nesred expressions like these occurringon both sulefi of the 'central' agent/patient arrow, as in:

[[[A] -; BJ -; C] --->(1) -; [E -; [PJJj.

~he o~ly reason why \\:e have not encountercd any formulae of this type so farIS th~t III the ones conSIdered so far there have only been four terms, and I..'ach(the Illdex, the artist, the recipient, and the prototype) has occurred anI\' one'Bu~ this restrict.ion is by no means stipulatcd. For inslance, the recipi;nt ea~eaSIly appear tWIce, once as 'the p-.ltron' and again as 'the public'. For instancelet us suppose that the index is Michelangelo's A1om, which, ever\'bo(h~kn?,,,s, was carved in order 10 memorialize Pope Julius II. This work i~dex~

MJchel.'lngelo~s agency, but he was not 'prime movcr'; indeed, it is fair to sayrhat thl~ carV11lg (and the 'slaves' which were to accompany it) expresses inpart, MIchelangelo's experience as a patient, working for great patrons, \~'howere ,:"ore powerful and consequential than (e\'en) he was. Moses (the prOto­tyr=) IS a m~taphor for the Pope? a~d Moses, in our mythology, is archetypallya ~ase-man. A formula for thIS mdex must therefore include the recipienttWICe; once [0 represent the parron who acted as prime mo\'er, eliciting .Miche­l~ngelo's a~ency as a carver, and secondly to represent the public who ,IreSImply patl~nts, awed and over,whelmed by Michelangelo's artistic agency,Ibr?ugh which they become sub)t.'C1: to the indirect (social/political) agency ofJ~h~ II: To express this, it is necessary to resort to superscript numberings todIstingUIsh the two types of recipient:

[[[[Recipient-A'] -+ Prowrype-A] ~ Artist-A] -+ Index-A] -7

Recipient-P'

\\ here R~'C!picnr-A' :::: Julius II, Protorype-A =: Moses, Artist-I\ =: .~ lichelangelo,and Reciplelll-P' =: rhe general public.

4·4· SOllie more GOlnple:r T,.ee-Strucl/ires: The Nail Fetish

In order to explore some of the complexities of trcc-srructures let us considera t) pc of image well known to visitor.; to ethnographic art museums· I referto th~ 'nail fetishes' of the Congo region of West Africa (Fig. ........./I)~ ThesesTarthng figures, anthropomorphic in form, arc inst;lntly identifiable because ofthe nails dri.vell infO their bodies, a violation of the notion of the (semi-s,lcred)artwork whIch greatly adds ro the 'aesthelic' ji-;ss/)// they provoke in WestcrnspectatOrs, who perhaps wonder what !llichclangclo's Davit! would look like~ivcn the same treatment. Howc\'er, the apparent rh)me bel\\et'n thcst: can'~mgs and \Vestern images of suffering and violation is fortuitous, and the actualnetwork or agency-relations which surrounded them in their original seering

60 The [m;o!lIfiolllJ!lhe {/ldex The lll'i)ollition oflhe Index 61

, h 't t'ng th,on tl," f:lci1c imag")' of victimization which is allIS muc more In eres I .. ""

that uninstructed aesthetes can extract from them. .According to the missionary anthropologisT Denncr~, who work.cd 1Il ~he

C"h the colonial «overnment was bustly engaged In root11lgongo ,It a tlme ,\, en ' b .' .. d' ,

'I" '1 'th, h,lief that rhe\' werc foumamhcads of native se 111011,Ollt nill diS les In '. .these images were essentially judicial in functlOll; they belong ,to t~c same

h 0' I'" I ",o'k~' which presided over legal procccdmgs In manycategory as t C Jut loa ".~ ~ . .," ". 'I' '[I," ,"mlicial mask I)Umshes those who he all oath; theparts 0 . est rio. net.. ... . . .... . '"

'I' 'I ' ,th :mihr f:IShlOll registers pr0\1llses and o,nhs, and pUnishesnal IctlS 1, In la erSl " ,

~!G. 4.4/2. The fetish as aninuex 01 cumulative agencyand ~s the visible knot lyingtogether an invisible shin ofsp~tio-!empor~l relalions

Inclex-A~ Recipient-P2Fetish • VictimI :~dsesioo!C1al

~~Ilyover

Suppiicant FetishReeipient-A1 fflokes (by driving Iinclex-p

tonall~

Muamba tree Brave hunterinAcausesdeathOf Inclex-P

Priest _ Muamba tree Brave hunter Wild animalsArtist-A outs down Index-P cau~th 01

A palaver Lmeetin[il is held, and it is decided whose Kulu LsoulJ it is that is to entcr intothe J'vl uamba tree and to presidc over the fetish to be made. ,1\ boy of great spirit, orelse, above all, a great and daring hunter, is l:hosen. Then they go into the bush and eallhis name. The Ngang-J LpriestJ <.:uts down the tree, and blood is said to gush forth. Afowl is killed and iL~ hlood min[ikd with the blood that they say comes from the tree. Thenamed one then dies, <.:ertainly within tell days. His life has heen sacrilieed for what theZinganga l:onsider the welfare of the people. They ~ay that the named one never failsto die. _. People pass before these felishes (Zinki<.:i Mbowu) calling on them Lo killth~m if they do, or have done, such and su<.:h a thing. Others go to them and insist llpolltheir killing so and so, who ha~ done, or is about to do them some fearful injury. Andas they swear and make their demand, a nail is driven into the fetish, and the palaverrbusincss] is settled so far as they ~n: <.:oneerned. The Kulll of the man whose life wassacrificed upon the clltting of the tree sees to the rest. (R. l)ennetL 1906: 93)

There are pronounced resonanl:es between this ritual sequence and The mak­ing of the allkishi figurines from the blood-exuding Mukula tree among theNdembu (see above, Sect_ 3.2). Structurally, the siLltation is rr[[Artist-A]---7lndex'-A] ---7 Recipient'-Al----') Index'-A] ---)0 [Recipient"-Pl But Lhis hardlydoes justice to the complexity of the situation, which is more perspil:uously shownin a tree diagram, as in Fig_ 4-4/2. This tree dill gram contains features not en­countcred before. First, in the formation ofrhe 'index' (the fetish) it is, by turns,passive (the tree whid1 is cut down), thcn active (the tree which metonymicallybrings about the death of the hunter whose name it bears), Then passive again(the fetish which has nails driven into it, 'attaching' the supplicant's requeststo it), Then, finally, active again (as it executes iTS judicial functions). The indexis doubly active because, in the process of its formation it has been doubly

those who contravene them. According to !Jennett, when such a judicial fetishwas to he made:

nritish i\lm;eum, London,FIG. " ...11. Nail relish figures li'om the Congo regiun of /Urica.iVUvlo2.1467_~lm;cc de 1'Ilommc, Paris. Kongo, Loungo rcg'on

62 Tllr Im;tllll/;f1/l oflilt indt'X The IlJ1'tJ/Ulion Ilflhr 111I10. 63

passive; it has the capacity to act (as a fetish), because il has heen acud upon, bothas a rrcc, and, simultant.:ously, as a hunter who dies 'for the welfare of the people'.It demonstrates, in Bloch's terms, 'rebounding violence' (1991). Secondly, thetree diagram bifurcates at the base; the nail fetish has nOI onc 'root-man' butIWO, the pri~1: and Ihe hunter, the fonner being responsible for the creation ofthe index as an 3ncfact, the latter being responsible for its efficacy, which isunderwrinen by the humer's cffi<:acy in slaughtering brame animals.

An instructed person, approaching such a fetish, does not sec a mere thing, aform, to which he Illay or may not rtospond aesthetically. Instead, what is st.'Cn isthe visible knot which ties LOgcthcr an invisible skein of relations, fanning out intosocial space and social time. 'fhese relations are not referred to symbolically, asif they could exist independemly of their manifestation in (hii'> particular form;for these relationi'> have produced this particular thing in its concrete, factual,presence; and it is because these relations exist(ed) that the fetish can exerciseits judil.;al role. However, I shall defer further discussion of this theme untila later chapter, dcvoted to the whole subject of idolatry. The purpose of thisdiscussion is only to show the 'involute' character of thc index, which mayobjectify a whole series of relations in a single visible fonn.

Having disparaged the Western response which sees the nail fetish as animage of surTering, let me now turn to another example in which the surTeringsof the index do genuinely connote the surTering of the prOlotype. Howc\'cr, the'work of art' I have in mind, though belonging to the West, has so far nOlentered the callon of Western art'. I refer to the 'Slashed' Rokeby VemiS, thework of a suffragette artist, Mary Ril.:hardson (and Velazquez). This work onlyexisted for a few monthi'>, before it was superseded by lhe 'restored Rol:ebyVenus' which can be seen in the National Gallery today (by Velazquez and theMuseum's picture-restoration staff). Fortunately, Mary Richardson's versionof the picture was phowgraphed, and Ihis image is reproduced by Freedberg(1989: +1I). Sec Fig. +..+13.

Frecdbcrg devotes a whole chapter of his work to the attacks made on im­ponant worb of art by so-called fanatics, such as Mary Richardson ('Slasher.Mary') who attacked thc Rok~by V~nlls with a kitchen knife in 1914, Frccdbergmakes the poim that, though the Gallery officials always cxpress incredulityand dismay after such attacks, ascribing them to insane malice, there alwaysdocs prove to be a strong religious or polilical motive prcscnt in the mind ofthe attackcr. For instance, it W:IS not by chance that the latest picture to heattacked in the National Gallery (in 1978) was Poussin's The Colden Callnorthat the attacker's Ihrusts were aimed at the Colden Calf itself. Nothing wasmade of this at the time, bue Freedberg argues, surely correctly, that this pic­(Ure was chosen precisely because it depicts idolatry in progress. The allae\.:was directed a~inst idolauy, and with re-olson, because in lhe National Gallery,even if we do not commit full-blown idolat'T), \\e do \'erge on it all the time,

F~(J_ .. ·413· Mr" Parlkh/1rS1 by Ahry Riehard~on: the RtJkeb.r VitiII>' by VeLizquez slashcd b} J\'l:iryR,chard.<;()n, Iljl'" SOI/TU: Thc :'-Iatiollal Gallery, London

Th~s atl'olck ~\'as consigned to the category of 'arbilrary' outrages committed byschizophrelllcs lx:cause to admit otherwise is to drop the defences wc crectbetween ourseh'cs and images which mo\'c us. Freedbcrg san; ,hat 'the ;con~clastic act is so frightening' because: •

~f opens realms ofpo\\er and fear that \Ie ma~ sense bur cannol tillite grasp. "'hen Ihelconoclas~ reacrs with ,\iulcnce to lhe imaJ:,'t: and \-ehemclllir and d(';lmaticall~ alleml)tS10 breal; IlS hof~ o~ f~lm or hcr, then \1 c begin fO hal'C some ~11SC of irs 1>Olcnlial-ifII"C do not pCrl.'Cll'l: II In the /lash oflight lhal hlinds us, finally, to its art. Butlhcsc darsII c hal'e become lllorc sophistil;;lted, and therehy morc confused. We a)]o\\, that ;tTl (.~n

?e trouhling. 100, nnd we come full circle. We have learned to lurn the troubling imal;cIllW someth\l1g lI'e e,m s,l!d)" call art. (l'rccdberg T9ill): ,1-25)

rn::cdberg makes a poimed cOlllr,lst between the ostensible violence of Illllchcontemp~rary Wester.n :1Tt, whieh smooth-talking crilics :tnd collel.:tors pT:liscto the skies (because U IS art) :lI1d Ihe shock-horror reaclions thai actual vio­lence ag;linst works of art provokes in rhe same quarters. But it is a moot

64 !'lie in1!o!lIlion ofille index jlz" Im:o!ution oflhe Index 65point whether the outsiders who, due to mental instability or strong politicalmotives, overcome the taboo against defacing masterpieccs in museums, arenot the more dcmonstrably in thrall to an. Such persons, as Freedberg puts it,are 'blinded', by art-blinded, that is, to the fact that art is not the real thing.Although, in the twentieth ccntury, such attitudcs can only really expressthemselvcs in acts of extreme deviance, they arc natural and basic, rather thanobscure and fanrastic. Art-destruction is art-making in reverse; but it has thesame basic conceptual structure. Iconoclasts exercise a type of 'artistic 'lgency'.

Let us examine the case of 'Slasher Alary' and the Rokeby Venus, the mostfamous act of iconoclasm of the 'modern' period, and moreO\'er, one not dueto obvious insanity but to clear ideological motivcs..I'vtary Richardson gave thisaccount of her action in 1914 'I have tried to destroy the picture of the mostbeautiful woman in mythological history as a protest against the governmentfor dcstroying !vlrs Pankhurst, the most beautiful character in modern history'(Freedberg H)89: 5(2). From this and from a later interview, it is quite clearthat Richardson equated the woman in the picture (Venus) with EmmelinePankhurst, and the 'sufferings' of the picture with Alrs Pankhurst's sufferingsin prison. In other words, this is an instance of volt sorcery (sec below) inrewrse; the sufferings of the victim cause a change in the appearance of therepresentation. Examining the photograph of the Rokeby Venus after the attack,we note that the deepest slash is aimed at the heart; Venus has been stabbedin the back-a very political way to die. In effect, Mary Richardson was anartist who produced a 'new', modern, Rokebv Venus, now a representationof Emmeline Pankhurst (standing for modern womanhood as Venus stoodfor mythological womanhood). The 'Slashed' Rokeby Vcnus by Richardson, is,without question, a more powerful image than the old one by Velazquez,though infinitely less aesthetic, because the image bears traces which testifydirectly to, rather than simply represent, the violence womcn endure, or believethey endure. The contrast between the supremely controlled and detachedagency exercised by Vel:izqnez in creating the painted image of Venus, and thefrenzied gestures of Richardson defacing the image so that its 'death' COITes­ponds to that of Pankhurst create the space in which the life of images andpersons meet and merge together. Richardson endowed the R()ke~V Venus witha life it never possessed before by 'killing' it and turning it into a beautifulcorpse. The restoration of the picture to its original condition, though ofcourse necessary and desirable, was also a means of re-erecting the barrierwhich prevents such images troubling us unduly, politically, sexually, or in anyother way.

The structure of the agent/patient relations surrounding Richardson's'Slasher!' RokeD)' Venus arc depicted in Fig. 4.4/4. Here we have a completeduplication of prototype, artist and index and recipient; the two prototypesare Venus and Mrs Pankhurst, the two artists arc Vcla:lque:l and MaryRichardson, lhe two indexes are the Rokeby Venus in its intact and slashed

---~---,-- ---The 'Slashed'Rokeby Venus, .. Roc - 1Inde.2_A J """s,won, ouu\/<, 1pt!!n1 .p (the art pulllrl:,~ t1egovera-nent, etc,)

Mary RiChardson l'iole<Ico .. The Rokeby Venus

Artist2~ ~nde.LP

Wrs PankhlJst~MafY Richardson VelazqlJ€!z~Venus

Prototype I_A "'Spffos Recipient2-p Artisll.A ,,",pic,s Prototype 2.p

Prison warders MfS PankhlJst ( .. ) Venus Artjst-P

Re~'pient 1.A VlJil>n<o "':'----7::...---------J .~~ProtolypO l_p Prototype 2_A

FI". ,H'!4, The 8h~red biographic~l spa<:es of persuns and images

state, ~'he two re.cipients arc iVlary Richardson and the outraged art public. Theresultmg Tree dIagram consequently bifurcates thou"h the two branch .." ,J.. J' ' D "",, soImp Icitly join again at the base, as a result of the quasi-identity between the'mythological' heroine, Venus, and her 'modern' counterpart, .Mrs Pankhurst-note also the imp.licit identity between Recipient-A (the prison warders,agents of tlle repressn·e government) and Recipient-P (the outraged public).

The OriKil/(l/ilJlIlJjlllt' lIlila 67

5

The Origination of the Index

5.1. Agel/c)1

The fact thai I have chosen to employ a single graphic symbol, an arrow, couldbe taken to imply that 'agency' has some quintessential, generic form, of whichthe "arious types of agency so faT mentioned arc species. This inference wouldbe incorrect; the ::agency :lITOW implies no panicular kind of agency, onl~ thepolarity of agent/paticm relations. I sct no limit whatsoe\"er to the type of'action' involved. Sometimes this action is psychologil..-:al; for example, the'action' of an index in impressing a spectator with its technical excellence, orarousing the spectator erotically; while at other times the action may be phys­ical, as happens, for instance, if the index is a holy icon whil;h I..-urcs therheumatism of the one who kisses it, rather than mere!} looks at iL Com'en­tional 'theories of art' arc mostly predicated on one, or a limited selection, of'kinds ofagency'. Thus, aesthetic thL''()fics of art are predicated on the idea thatartists arc cxclusi\oeJ) aesthetic agents, who produce works of art \\ hich mani­fest their acsthctie intcntions, and that these intentions arc communicated tothe public which \'iews their worl.s in the light of approximately the same setof aesthetic intentions, \'icariously emertained. In an idL"'31 an world, suchmight indeed be the case, and nobody would ha\'e recourse to works ofart withanything in mind except thc garneting of aesthetic experiences, and ccrtJinlynot in the hope of being cured of rheumatism. Semiologic or interpretJIivethL'()riL'S ofart assumC that wo,'ks ofart are vehicles of meaning (signs, symbols)which spectaTors havc to ,!L't-"Ode on the basis of their familiarity with the semi­olugical r>yr>tem ur>ed by the anist 10 cnwde the meanings they contain. I do notdenv th;lt works of art arc sometimes intended and receivcd as objects of aes­thetic apprecialion, and that it is sometimes t.he case that works of .ITt functionsemioticlllly, but I specifically reject the notion that they always do.

The kind of ;lgency exen;ised in the vicinity of works of art varies ~onsider­

ably, depending all a number of coTHextual filCtors. In gross terms, 1t may besupposed that whatever type of action a person may perform vis-a-vis ;lI1ocherperr>OIl, may be performed also by a work of art, in the realms of the imagina­tion ifnol in reality-not that we are always in a position 10 decide wlmt is 'real'and what is not. ~rhe anthropology of art, to reiterate, is just :lIlthropology

y itself, eXCept that it deals with those situations in which there is:l11 'index ofagency' which is normally some kind of artefact.

There are instances in which the index may at:tually be a perSOll. A case inpoint is possession by the deity. In Nepal (and elsewhere in the Hindu world)young girls are worshipped as the guddess Durga (Allen [976). The kumari(living goddess) is an indcx of lhe deity, and it is impossible to distinguishbetween the cuil of the living goddess during the festival, and the cult of idols(//Iur/i) of the samc goddess;lt other times. The young maiden is also a nll/fli­an image of the goddess-but a li\ ing one.' The living idol of the goddess findsa secular counlerpart in Western contemporary performance an. The actressTilda Swinton was recently exhibited in a g:ll1ery as a Ii\'ing (immobile, sleep­ing) work of an, and rhere are numerous simi1:lr examples.

Any dramatic performancc il1\'olvcs one pcrson (an actor) serving as theindex of another; that is, the character being enacted. Acting is, in general, aform of representational an, and ir em be expressed by an identical formula:

[[LArtist-Al ~ Prototype-Al ~ Index-Al-+ Rccipient~P

where Artist::: playwright, Prot'Ot~pe = character, Index::::: aClOr, and Recipient::: audience. And ofcourse, the dramarurgical situation may be modified at will,just as the situation vis-a-vis works ofart in the form ofartefacts may be. Thus,in ostensibly 'inwluntary' dr:lmatic roles (such as J>O!'scssion by the deity), thefonnula would inYoh'e the indcx in 'patient' position:

ProlOtype-A -+ llndex-P ~ [Rccipiem-P] 1where the Prototype is the deifY, the Index is the possessed shaman, and theRecipient is the congrcg:ltion.

These remarks will be sufficicnt to indic,,ue that there is seamless continu­i~- belween motles of artistic acrion which in\'oh·e 'performance' and thosewhich arc mediated via artefacts. The distinction has no theoretical significance.Eyery artefact is a 'performance' in that it motivates the abduction of itscoming-into-bcing in the \lorld. Any object lhat one encounters in rhe worldilH"ites the question 'how did this rhing get to be here?' Mostly, the answersto such questions ;lre so mken for grantcd as nor to play any part in one'sconsciow, mental life (hut somewhere or othcr in one's psyche, therc mustbe a de\,ice which identifies the familiar (IS familiar). Only gmlogists, who aretraincd to do so, ask, when they sec a mountain range, how that came intobeing. But with artefacts, which arc the product of types of agenc), w11ich wepossess generically, the sitwlIion is often very different, and we do indeed con­sciously attend to their originr>. This means playing out their origin-storiesmentally, rceonstructing their histories as a sequence of actions performed byanother agent (the nrtist), or;l 111ultimde of agents, in the instancc of collect­ive works of :lrt such as cathedrals. \Vc cannot, in general, takc up a point ofview on the origination of;ln anclact which is lhe poim of view of the artdilctitself. Our natllral point of vant;\ge is lhat of the originating person, the artist,

, For 'llor£ on lh~ 1m",,,,, cull of li"ill!!" goddcsses "'"'c helm,. &£1. 7.'2.

68 The Origwalion of'''r JIII/t.r

because we, also, aTC persons. We imagine the origination of a painting fromthe vicariollsly entertained standpoint of the painter, nor (as we might) fromthe stanupoiJll of the paim or canvas. Only when lhe anefact acrually is a per­son (as in the case of pos5(.'SSion by the di,-inily, jWit mentioned) might we adoptthe index's point of vicw, but this invoh'cs :m implicit refusal of the scenarioof possession, in thal the possessed person is temporarily in abeyance, havinglx,"COme the "chide of the personhood or the di\·inity. So as a general rule, it is['liT to sa\' rh:lt indexes, from the spectators' point of view, only mediate person­hood TatilCT than pos.<;ess il intrinsically. However, the personhood of the artist,the prototype, or the recipient can fully im-est the index in artef.1ctual (onn, sothat to ,Ill intents and purposes it becomes a person, or at least a panial person.Ir is a congealed residue of performalll.."C and agency in object-form, rhroughwhich access to other persons am be attained, and via which their agency (:anbe communicated.

5.2. Captivation

Thcoretically, there is no limit 10 the unu of agency which can be mediated byindexes, but it would be disingenuous on my part to sUbrgest thaI I uo nOI artachmore priority to certain types of agency than to others. Where indexes are veryrecognizably worl.:s of art, made with technical expertise ,md imagination of ahigh order, which exploil the intrinsic mechanisms of visual cognition wilhsubtle psychological insight, then we arc dealing with a Glllonicai furm of art­istic agency which deserves specific discussion..\rlany indexcs are crude anduninteresting artefacts, whose importance rcsts !>Oldy on Iheir mediatory func­tion in a parrieular social context-for instance, the figurines used in Africandivination--and while these certainly fall within the scope of the anlhropo­logy of art they have no significance as 'works of art', because nobody anendsto their making as a parricularl}' salienl fearure of rheir agency. It is other­wise with :lTtefacts which announce lhemscln:s as miraculous creations. The'coming into being' of these objecls is explicitly attended to, because theirpower partly rests on the fact thar lheir origination is inexplicable except as amagical, supematurnl, occurrence.

In cultures which produce an at all, most 'Hlults (of the right gender, wherearl production is gendered) ha\'e, al some st.lge, auempted to originate worksof art, at leasr of a trivial or ephemer.tl nature. Speciali7.cd art production, associal praelice, implies that mOSl adults are either f.1iled, or rehui\'ely unsuc­cessful aniSls· onlv a few lalented individuals, and{or individuals who receiveinstitutional ~neo~ragemenl, specialize in the production of really fine work.They are the artists. The biographical prohability Ihat passive recipients ofworks of art h:l\·e some background praclical experience of the art-makingproct.·ss-whether this is fonnal school instruction in 'an' in the \Vest, orjust childhood experimentation with whittling Slicks amI tracing patterns in

the dust in. non-\.\:estern cuhures-ensures thaI Ihe reception of a work ofart occurs m the Ilghl of Ihe pQss;hili~)r th'lI the recipient could, u:chnical1y,approach the same task of art-making, himself or herself.

Thus, parI of Illy experience as a recipient of Vermeer's Lllremflka is theco~ne.mplation .of th.e possibilily thaI I, not Vermeer, could have produced this~3111tJ~g-notm thIS world, I hasten to add, but in some olher 'possible world'III whIch I would be a much bener painter than I actuallv am. At the sametime, I am acutely aware of the collnterfactuality of this ;pparently feasiblew?rld: e\en ~hough I know (generically) how to mix paint, and I can drawalter my fashulll, I also know that I could not even produce a decent COP\' ofT/,~ UlCmltl~f!r, let alone originate a comparable maslerpiece (The Seamstressby ,\Ifrcd ~~n. Gcll). Gazing at the picture, my jaw drops, in admiration­and defeat. 1 hIS defeat is, however, profitable to me also to the extent that in~lent.ally. retracing Vermeer's origination of his pictur~, the technical andImagl.n~t1\'e performance which culminated in rhe finished work, I do manage,exe~clsmg .such powers as I possess, to attain a certain poinr, before I breako~ Tn ?e~\'lldermelH and can follow Vermeer no longer through the maze ofhIS artistiC agency. Up to a poim, I can be Vermeer 1 can idelllify with hisartistic pT(x:cd.ure and see his picture, vicariously, as'a product of ~y bodilyeng-.lge.ment ~Vlth the world and with the materials artists manipulate. !Jut onL"Crhe l?omt o~ Illco.mmensurability is reached, the point at which it is no longerPOSSible to IdentIfy Vermeer's agenc)' with my own, then I ,un left suspendedbe~ween two worlds; the world in which I ordinarily live, in which objects haveratIOnal explanations and knowable origins, and the world adumbr:lIed in thepict~re, w.hich deftats explanation. Between thtS(: two worlds, I am trapped ina IOgI.ca.l blt1d; ~ must accept that Vermeer's painting is part of'my' world-forhere Ir IS, physIC'J.lIy before n1l"--while at the same time it cannot belong to rhis\\'?rl~ b.CC3usc I only know this world through my expcrience of being an agelll\\'Ilhm It, and I cannot aehie\'C the necessary congruence between my expcri­en=~ (~f ~gcncy .and. the agency (Vermeer's) which originated Ihe painting.

. ~ hIS IS ~pt~vau(}n, the primordial kind of artislic agem:y. It is folr fromdifficult to CIte II1st3nel'S of the deployment ofcaptivation in practical contextso~her I.han Ihe an gall.ery. In all eal'lier publication (Gel I 1992h), 1 discussedt e efficacy of Trobn,md canoe prow-boards as psychological weapons inthe context of. overseas Kula exchanges (Fig. S.2{r). These boards arc richh.ean'ed and palllted, and they are the first thing lhat rhe Trobrianders' o\"er~seas exchange-partners get to see when the Trobriand flotilla arrives on theirs!lores, b~fore. exchange operations get under way. The purpose oflhese beau­tiful ~-arvmgs IS 10 (~emornliz~ the opposition, SO thai Ihe) will lose the capac:itvto ~n\e hard .b~lrg:uns or.r~slsl .the Trobrianders' blandishments and plausibl~f,1Iseh(~s. !,e~ther the 1 robnan~ers .nor their exchange-partners Operale aealcgor) of aTl as such; from. theIr pOl.Ill.of ,""iew the l:ffi,,-acy of these boardssIems from the powerful m:lglL"31 aSSOC1;lllons lhe\' h~'·Q A I d'• ".... prow- l(Xlr IS an

index of superior artistic agency, ;md it demoralizes the opposition becausethey cannot mentally encompass the process of its origination, just as r cannotmcnt>llly encompass the origination of a Vermeer.

Hut the paradox of the incommensurability of creative powers is under­stood in different terms; the artistic agency on display is magical, the resultof the artist's ingestion of a magical tradition and substances which commun­icate carving skill. _~·,tlagic, in the Trobriands or anywhere else, is not 'normal'even though magic may be in everyday usc for a variety of purposes. TheTrohriandcrs live in exactly the same world as we do and operate exactly rhesame conceptual categories so filr as 'ordinary' cause and effecl' goes. \Vithmagic, it is different; magic produces exrraordinary dfects by means which donot articulate to the agent's 'normal' sense ofseH~ embodiedness, agency, andbeing in the world. Thus, the fact that the boards arc said by the Trobriandersto be 'magically' efficacious in demoralizing the opposition is just a tran­scription of the experience of captivation into the language of magical causa­tion, which we arc also tempted LO use to describe the same type of 'uncanny'sensation which great works of Western art produce in \Vestern spectators.Artistic agency, especially of the virtuoso character so obviously present inTrobriand carving, is socially efficacious because it establishes an inequalitybetween the agency responsible for the production of the work of art, and thespectators; in the Trohriands this inequality is attributed to superior magic;in the \Vest, to artistic inspiration or genius. Neither 'explanation' is reallyexplanatury, each only serves to register the disparity of powers between artistsand spectators. This fundamental inequality of powers is carried over inLO thewider social transactions within which the art object is embedded and mobi­lized; the kula exchange. The Trobrianders, if their artistic magic works as itshould, are one-up before the exchanges even begin; having demonstrated themagical poLency Lhey pos~ess in the artistic domain, the implication is thattheir exchange-facilitation magic is equally effective. They arc irresistible, andtheir exchange-partners will find themselves, willy-nilly, disgorging their bestvaluables withouL demur.

I have no wish to recapitulate any more of the points I made on the basis ofthe Trobriand example in my previous paper (Gell I 992h). There are some,ldditional comments though, which may serve LO relate the notion of captiva­tion as a form of artistic agency to the general argument being advanced here.Captivatiun or fascination-the demoralization produced by the spectacle ofunimaginable virtuosity-ensues from lhe spectator becoming trapped withinthe index because the index embodies agency which is essentially indecipher­able. Panly this comes from the spectator's inability mentally to rehearse theorigination of the index from the point of view of the originator, the anist. The'blockage' in cognition arises at the point when the spectator cannot followthe sequcncc of steps in thc artist's 'performance' (the 'performance' whichis objectively congealed in the finishcd work). Thc raw matcrial of the work

70 The Origina/ioll oIlht' Jndex

I'IG. 5.zlI. Trohriand can0~ prow_buard. S"ura: Shifl~;' Campbell

1'lie Origil/alion of/he Intlex 71

J2 Tilt Or/xina/ion of/he Inde.\'

(wood) L-an be infern...d from the finished producl, :lnd the basic technical steps-­carving and painting; bur not lhe crit;L-a1 path of specific H.'chn;I,:01I processesalong the way which aCLu:llly Cnttt the transformation from ra.", .male~i~1 to

finished producT. In other words, it is the complexity of the arllstlc deClSIOIl­making process (gcner:Hc and [cst) which defeats specrawrial n:capitulation.This relates to the earlier discussion of artistic decision-making and proprio­ception under the heading Artist-A ----7 Artist-P (3OO\'C, Sect. 3. 11 ).

But l..':lptiYation has other sources as well. The emphasis 1 have been placingon artistic 3<Tcncy has cmpiric:.IJ supPOrt. in thal we in [he West do hero-. . .worship artists, and the Trobrianc:lers do place inordinate \'alue on artistic

Prowess thou...h thcv associate this with possession of magie-ill resources,, . ,rather than genius. UUI there are many !}'Pcs of abduclions of agency fromthe index, and the abduction of origination in artiSlic agency is only one ofthem. Perhaps, though I do not think so, I place so much emphasis on captiva­tion through artistic virruosity as a mode of agency because I myself am a'Sunda\' painter', and I consequently have a propensity w im'lg:ine, when I seca Verm'eer, that I am Vermeer, painting: thus and tlms. The tragedy of amateurartists like me is lhat because they can partially recapiru1:lte The performancesof their artistic idols, they know much better than non-artists how abjectlythe" lail to achieve true virtuosity. Others, whose own pr,lctical acquaintancewitil wielding pencils and brushes is lost in the mists of childhood, or has beenexpunged with other degrading: adolcscenr experiences, may never experi­ence the impulse to measure themselves against the gn:at artists. On the otherhand, I ha\·e ne\'er e\'en attempted to play the cello, yet I think I respond to

Rostropo\'ich recordings in l.erms of imaginary manipulations of an imagin~ry

cello, with approximatc1~ the same results as my imaginary efforts to pamtVermeer's pictures; so I do not think immediate practic:ll experiencc as a'failed' yirtuoso is noct's.'.an' in order to become t:aptivatcd by \irtuosity,

Whatenr the intcrsubj<.';"ti\'e validi!}' of thc analysis of cptivation in termsof indecipherable agency may he, I would agree that captivation can occur inother ways. l\'luch art criticism downpla~'s virtuosil~ and indt:cd artistic agencyaltogethL~, and concentrates on the visual-acsthetic properties of art obj~ctsmuch as if thc\' had come into being by themselvcs, quite withollt the phYSicalintervention a"f am' anisl. The art object 'in itself' is lhe focus of attention,rather than the p;occss of its origination through the bodily aClivities of an,Irtisl. Leaving artistic agency ro one side therefore, I shall now tUTn the .spot­light on the naturc of the capti"ati~ll1 exerted b.y the~ndex ill and of Itself,rather than as the outcome of the pT10r agency 01 all arus\.

6

The Critique of the Index

6.1. On Decorative Art

Since thL"iC reflections on art arc directed in the main at the existing anthrl)­pology-of-art confralernity, I can make a starr-and perhaps make amcnds­by considering, first of all, a classical theme in the anthropology-of-art lit­eratu~e, namel), the significance of 'geometric' (i.e. non-representational, ormargmally representJtional) deeoratiyc dcsigns.."'lost 'Western' art-theory isabout rcpresent,ltional arr, and so is most of the philosophy of art, in so far asthese two may be distinguished. But it is fair to say thal most of the ,1rt in thecollections held in ethnolngical museums-if not on display there-is 'decorat­ive' art, usually applied To artefacts such as pOTS, m.lts, and so forth. Manv ofthe morc intereSTing studies which lla\'c been produced by anthropologistsconccrned with art (e.g, Kacppler Il)78; Hanson 1983; Price and Price 1980)ha\'e 10 do with lhis kind ofart. Cert.1inly lhe most massi\'e compilation ofdatafor the study of non-Western art (Carpenter and Schuster 1986) deals primar­ily with ostensibly decorati\'e art, though these aurhors' approach is predicatedon the idea that apparently dccomtive forms ha\'e universal symbolic meaning.

There i!> another I'cason for commencing the substami\e discussion of 'theanthropo~'ofart' with a umsideration ofdccoratl\'c art, and that is the elimin­ation of a form of gender bias which is prcvalellt in much of the anthropology­of-an literature (this work included), which pays most attention to contextsofart production dominaled b} mcn, such as gender-c.'l:clusivc male cult riruals,It is in the context of male cults, or cult aClh ities dominated by men, thatart production occurs in forms which bC'.lr the most imnu:diatc comparisonwith Western 'fine' art; hUI Ihis is no reason for always gh-ing this typc of arlproduction pride tlf pbce, anal~ tically. 'fhc development of (Western) abstractart during the twentieth centur}, and the simult;tneous rise of 'design' 10 a St:llUSrivalling, or c\ cn exceeding, the prestige of fine art, indicates a change inattitude which Illay also be extended to the non-representational an and/or'design' produccd by non-Western artists. Many, even most, decorative artistsworld~wide arc women, because of the frC(lllency with which thc division ofhlbourIn agr:lri:mfsubsistellcc SOcieties assigns textile prod llction, potlery, baskelTY, andthe like, to women. Which is not to say that there arc nOt. male decorative artistsas well. The advant,lge of beginning with dccorari\'c art, hO\\Tver, is th:ll weare, so to speak, in :1 neutr:lllerrain, not one rivcn with \ iolent ideological andinstitutional Ttlssles, as is the case wilh much high-slalus rilual art.

74 111( Cnl;ql/{' oflh~ bl(J~xTk~ Cril;qu~u!fl,r JIlI'~:r 75

6.2. AJJOrltmelll

r'G_ 6.~'1. Jntmul Jimc-wmaint'r,>. SO/lrt~: Thc !'I'lu~um of Archatoology and AmhropoJog},C.:lmbndgc 35-8... 35-91, 30 37-1-, 35-8z. 35-8&

:Jllu~~d to before (w.hich. is the psychological aspect of the anthropology of arr)cOflJoms a theory of SOCial efficacy with considemlions which if not :lestheticare definitely eogniti\'e in nature. because cognition and socia'lilv arc one. '

Hm\ C".m such cognitive considerations be integrated into the theoreticalfrar:n~work advance~ hitherr~-Ih~ four basic terms, index, arrist, protot~ pe,reCipient, age~t/patlenl rdatlOnslllps, and so on? On the assumption that,where dcc?ratlve ~atterns ,lre concerned, there either is no prototype, or theprotol):pe IS not salient, the 'pattern' and the index are one and the same whichwoul~ Imply that the \\hole discussion lIQuId have 10 be carried OUI i~ termsof arr~st, Index, ,and recipient. This seems a rather limited set of concepts forha.ndl1ng compllcaled eogniti,-e, e\en possible acsrhclic, prohlems. How maythiS be accomplished?

Decorative patterns applicd to artefacts attach pcople to things, and to thesocial project~ those things cntail. A child may morc rc,ldily be induced to goto bed-which children are on'en disinclined 10 do-if the lx.'1.I in qucslion hassht.'Cts and a pillowcase richly embellisht.-d with spaceships, dinosaurs, or cwnpolka dOlS, be they sufficiently jolly and attracti\'c. The dCl;or.uion ofobjel;t.. isa component of a social technology, which 1 have elsewhere called thc techno­logy of enchantment (Gell It)/)2h). This psychologicaltcchnology encouragesand suslains the motivations necessitated by social life. The world is filled wilhdccor.lted objects beC:lUSC decOr.llion is often essential to rhe psychologicalfun<:tionalit), of artefacts, which cannot be dissociated from the other typesof functionality they possess. nOlably their practical, or social function:lIity.UndecoraTed children's bedsheets would be less functional in conferring pro­tection and comfort during slecp than decorated oncs because children wouldbe Icss inclined t,o sleep in them. They would be less socially functional, be­cause comfortable, protected, sleeping infants are a prime objectin pursuedby pa.rent<;.ln other words, the distinclion we make between 'mcre' decorationand function is unwarranted; decoralion is intrinsically functional, or else it<;prescm:e would be incxplicable.

Any bamboo tube, or container of suitable size and shape, could do duty asa lime-containcr, such as the ones I illlL<;trate in Fig. 6.2/1 from the lall11ul (NewGuinea). Bur a pL1in container would hardly be as functional, giycn that aman's lime-<:ontainer, in the COnlexl of labIml social life, is a mosl importantindex of its owner's personhood. For instance, rattling Ihe lime-slick in thelime-conTainer is a mC;lllS of communi<.:.lting passionat.e emoTion in oratoricalperformances (Bateson '973; Forge 1973). The dCl.:oration. which is distinct­iye, binds the Lime-container 10 it.... owner in a most intimalC fashion; it is lessa possession than a prosthesis, a bodily organ al;quired yia manufaL"ture andexchange rather than by biological growth. The ohjectification of personhoodin arrefacts amI exchange items is a familiar anthropological theme, on whichit is unnecessary tu dwel1 (e.g..IVlunl1 1973, 1986; M. Stnlthcrn 1988). ]]o\\,cver,anthropology has yet to thl-'Orize the nature oflhe attachment between persollsand Ihings mediated b~ surface decoration, and th,lt, rather than the prO\ isionof a more cthnographic context, is the task before us. Examining this gourdcontainer we are able lO sec thin it is dl-'COraled with beautiful p:illcrns. formedfrom mo;ifs that do not ob, iously resemble re,ll-world objects. The gourd'sdecoration is a free exercise in thc deployment of curves. ovals, ami spiralsand circle.<;, in synllnelrit.-al or rcperirivc arrangements. So far, I have not saidanything about such ostensibly 'acSlhctic' features of the indc.'I: as symmetry,elegance, etc. and it might ha\ c been assumed Ihat 1 ne,'cr intended to, havingruled oul the 'aesthetic' approach to the work ofarr in ad\·ance. This would bea mistake, ho\\,e\'cl", sincc the 'technology of enchanrmelll' approach which I

76 Thr Criliqut offhr IIIdtx Thr Critique ofthe b,Jrx 77

6.3. Deromtive P"llem = JlIdex-A -----711ldex-P

In Section 3.10 above, ] established the idc-J. that among the abductions ofagency we derive from the index arc agent/patient relationships between partsof the index ;;is-a-vis other parts of the index. Thus, in an abstract picture, onepatch of colour seems [Q be 'pushing against' another patch of colour, eventhough there are no objects in Lhe external world with which we call idenrifythese patches of colour. The causal interaction we percci\·c is internal to theindex itself. Part-to-pan 'causal imcr.lcl'ion' internal to Ihe index is the basisof 'decor:llive' art, which is just another name for abstract art. Decorative artinvolving the use of 'patterns' exploits the particularly (visually) salient part­to-part relationships produced by the rcpelitioll and symmelrical arr.lngementof motifs.

The application of a decorative pattern [Q an artefact multiplies the numberof its parts and the density of their internal relationships. The lime-comainersin Fig. 6.2/1, were they lacking in decoration, would be simple tubular formslacking in salient part-to-whole relationships. As it is, many separat.e pans maybe distinguished, and relationships between thcse parts. Each of these pansmay be considercd a subordinate index within the index as a wholc.

Parts of inde.us com'ey agency JUSt as indexes as wholes do. In the case ofdecorative art as opposed to representational an, it is not the di"crsity of theparts that communicates significant artistic agency, but their disposition withrespect to one :l.OOCher. In other words, the parts of the indcx exert causalinfluence over onc another and testify to thc agency of the index as:l whole, inthat it is in the disposition of the part." of the index that the artist's agency isprimarily madc apparent. We are, indeed, accustomed to speak of decoratedsurfaces (as well as representational indexes with a lor of physkal action inthcm) as 'animated'. This idiom reflects the fact that the motifs in decorath'can oftcn do seem [0 be engaged in a mazy dance in which our eyes becomcreadily lost. We need a formula which l:aptures the inherent agency in decor:u­ive forms, forms which do nOL simply refer [0 (represenL) agency in the extcr­nal world, hut which produce agency in r11e physical body of the index itself,so that it becomes a 'Ihing thing' without recourse to the imitation of any li,'­ing thing. DI."COrat'ion makes objects comc ali\'c in a non-representational way.Since this feat is accomplished \·ia pan-to--wholc relationships internal to thcindex, the decor.ltive index could he I'cprescnted as in Fig. 6.3/1 or:

[[rlndex-A motif"'''] --+ Index_Af"",iwIHlI<J --+ Index-A whole]~Recipient-P.

The root of a pattern is the marif, which enters into relationships withneighbouring motif.", relations which animate the inde.x as a whole, Howe\'er,we can dcscribe the rdevalll part-to-part and part-to-\\ hole relationships inpauerns much more precisely. Patterns can be distinguished from all mhcr

~Index .. Ro!l..'lp,,",nt

(lm.tc~) ~1,":C:::jf• (m.tn~' steps)

riG. 6.3/1. The hicr.:u·chiOlI Siructureof thc dcconllil'c indt:x

!I~dcxes by virtue of the fact that they ha vc salient visual properties of rcpetil_uenes." and s)'mmetr)'.' It \\ould he wrong to imamne that bed " . . .,. cause symmetry~n. repe~ltlOn are mat~em:ItlClI properties of forms, that it is not these proertlCS which most re-ddlly promkc the illusion-if it is onll' thot-of' l>­ca r,' h '. .. IlllmancO[.~sa It) In t e mdcx. :'\'othmg could be more animated than the t" II '

(tl h " ) ,. ' I h '. esse atlonsno patterns uCYlse( \' IslamIC decoratIve 'Irti""'I''''n'' b k 'II '"1'1 .r·'· <, "... " ~ u 00 I ummators.

. ~~ ~e IglOll~ly I~ll~osed ban on the r~pr~sentationofli\'ing forms only served,It secms, to IllSplre ever more eITecuve mdueemenrs to -pt,' t' b ' I, 'fi h '. ...... \'a Ion y VISWtallT ce, I e non-mimetic appearance of animal'ion.

6-4- Symmetry ami the Appearance ofAnimation

I-Iow.do mat.hemalical relations induce animacy in this way? The actual math­~matleal ~)a.:"IS of pattcrned fOl"ms is not hard to grasp conceptually, though it~s more dlffi~ult to a.pply. i\lI patterns are variations on only four 'rigid motionsIn the plane, to whleb rcpeated motif... can be sub)'ect~' ~I~.e. () fl'( . " cu. n....... are, I re ce-tlO.n, 2),translauon,.(3) r~ta~~on, and (4) glide reflection. These four motions~rc sho\\n a~d e.x~lallled m Fig. 6.411. The simplest type ofdceorati\'e panemIS the one-dimensional band pattern of the type shown in Fig. 6.4/2. This pal­tern (a ?rcek key) consists of the sUCcessive 'translittions' of a single motifal~ng a .lllle, It Seems .to move, because we 'read' it as we would a line of tcxt,b) movlI.lg Our anelltlon from motif to motif, which we must do in order to

otJ.sen'e ItS s.ymmerry, and h~nce. irs patterned-ness. The proccs~ of 'seeing'thiS pattern IS a matter of rl.'glstermg the fact Ihat motif A is identical to andto. the left of~ B, and C and 0, and so on.' Because the process of perceiving;~~s pat~ern l1l\"ot\·e~ ll1e~tally translating thc motif to the right, so as co lay

motif down on Its nClghbour ami register the congruence bctwecn them,

The Crilil/ll" ojlhe il/dex Tlte Crillqllt· of/he index 79

l.REFLECTION

2,TRANSLATION

'ROTATION n~4.GLlOE REFLECTION

A B'-'-" C .-.-.. etc.

1'1(.~. (1.412. The linear repetition of a singkmol if: simple raltcrn

PIG. (qIT. The basic m01iom ofpattemformation

dynamic property of the object being perceived~that still docs not explainwh~' patterned-ness is such a common property of artefacts, especially objectswhICh are pcrsotlill possessiuns and utensils, such as clothes, vessels, and so on.In order to understand this, one has [0 consider rather more complicated andrealistic examples. MosT significantly patterned objects and utensils are notdecora~e.d with 'simple' patterns, but with subtle and complicated ones, oftencompnsmg a great many motifs deployed in two dimensions simultaneousl\'and involving more recondite motions in the plane than simple translatio~:Sheer complexity, involution, and the simultaneous suggestion ofa great mallVformal relationships between motifs is a dJaracteristic of decorative art in gen~era!. One cannot understand decorative art by generalil,ing from simple, easilyil~tctv.reted, examples, because the Ie/os of decorative art lies in the opposit~dIrectIOn, towards the complex, the ambiguous, and the multitudinous.

Complicated patterns lie on the borderline between visual 'textures' and'shapes'. Gibson, in his accounr of visual perception in naturalistic (or 'ecolo­gical' contexts), makes ,I sharp distinction between the perception of textureswhi~h are structured surfaces (the sky, a green lawn, a brick wall) and the per~ceptlOn of shapes (a ball, ,1 rod, a cup, etc.). Textures :lre perceived hierarchic­ally constituted out of components (patches of white and blue, blades of grass,individual bricks, etc.) but these are not picked out indiridualIv- the\' are nor,lttended to as objects bur as components of a surface. \Ve can r~~pon·d to pat­terns as undifferentiated textures, and oftcn do, but not when we attend tothem as patterns. Howcver, when we do this, we cannot entirelv abstract themas 'shapes' either; they retain their essential 'texture' characteristic of internalhierarchy, division into motifs, blocs of motifs, and vague fringes.

agency and motion seems to inhere in the motifs themselves. The projection,or externalization, of the agency involved in perception (the perceptual act)into the thing perceived is, cognitivel}' speaking, the source of its animation. Itis unhelpful to describe the animation inherent ill patterns as an illusion, as ifit were some species of mistake. There is no mistake involved in describing thesun as 'moving' through the sky. Actually, it is the observer who is moving, notthe sun, but the sun's movement is not a purely subjective phenomenon, likea dream. Similarly, we do not dream that the constituent motifs in patterns,move, because these movements stem ultimately from the real movcments ofour bodies and perceptual organs, scanning the cnvironment. .

Psvchological experiments on apparent motion and apparent causality (suchas th~ famo~s experimenrs of lVlichotte) reveal how ready human subje~ts ar.eto attributc motion and mUTual causal interacTion to the harest of stlmult.Nevertheless, even if one can provide a cognitive exphmation for the animatedappearance of panerncd forms-as a phenomenon ofapparent mori~n wherebythe dynamic aspect of the act of perception is subjeerively expeneneed as a

6·5· Complex Patterns

Contrast, for instance, the Greek key (Fig. (1.4/2) with the design in Fig. 6.5/J,which is only a step up in complexity. It is quite hard, if one has not actuallytraced it out, to sec how it is orff<mized. \Ve perceive it simultaneouslv :IS a tex·­ture and as an arrangement of shapes (but precisely what arrangemen~ is harderto say). Still more beguiling arc patterns which reverse figure and ground,as in Fig. 6.5/2. This is a 'change' (colour-reversal) pattern based on glidereflcction along two parallel axes, along one of which the axe-like motif glidesonto a same-colour motif (white on white) and along the other of which themotif glides onto an opposite colour motif ofthe same shape (white onto black).It is one thing to be informed that the pattern in Fig. 6.5/2 is based on this par­ticular m:lthem:ltical rebtionship, but ir is vcry difficult 10 menrally followthrough the two types of glide reflection, so ['hat one can actually project therel;uionships onto the pattern.

In practice (and without the assistance of \V:lshbufll and Crowe), 1 thinkwhat happens is approximately as follows; what we do is single out, initially,:l

Ril Th., Griliqlle of/he 1m/ex The C,itiljllt· ofti,e Index

L'3ndidate motif, approximatcl)', lhe axe-like shape (cr. Fig. 6.5/2) which wespontaneously Sl.'C as being repeated (by translation) in rows :md columns, andwe may also be dimly aware that lhe 'ground' on whkh this shape is laid hasthe same shapes inscribed on it, buT in rcycrse colour. What, however, is likely(0 defeat us is seeing the relation between the figure :md the ground, and therelation of the lcliwards-pointing and rightwards-poinring QCCllrrences of Ollf

candidate 'basic' motif. So in fact we just mentally resign ourselves to JUSl/1li1

quite lIIlfler5talldill.~ these complex relationships, we write them in as 'beyondour ken'. We expericnc<.' this as a kind of pleasurable fruslr.uion; we are dnl\vnimo the pallcrn and held inside it, impaled, as it wcre, on its bristling hooksand spines. 'I'his paltern is a mind-tmp (cr. Gel! 1996), wc arc hookcd, and thiscauses us to relate in a certain way to the artefact which the pattern embellishes.

6,6, Complex Pllllems (H 'Unfinished Business'

Pauerns, by their multiplicity and the difficulty we ha\'c in grJ.sping theirmathcmatical or geometrical basis by mere visual inspection, gener:!te rcllltion­ships over lillie between persons and things, because what they present 10 themind is, cognitively speaking, always 'unfinished business'. Who, possessedof an intricate oriental carpet, c.1Il say that they have entirely come 10 gl'ipswith its pattern; yet how often the c}e rest's on it and singles out no\\' this rcl:l­lioll, this symmetr), now that. The process can continue interminably; thepattern is inexhaustible, the relationship benn:en c;lrpcl .lOd oWller. for life.i\mhropologists ha\'e long recognized that soci:ll rebtionships, to endure o\cr

rime, have to be founded on 'unfinished business'. The eSSence of exchange,as a hinding social force, is the dcl:ly, or lag, between transactions which, ifthe exchange relation is to endure, should ne"er rcsult in perfect reciprocation,but al\'<ays in some renewed, residual. imbalance. So it is with patterns; theyslow perception down, or e\'en hah it, so that the decorated ohjl.'Ct is neyer fullypossessed at all, but is always in the process of bet--oming possessed. This, I:argue, scI." up a biographical relation-an unfinished exchal1ge--between thedecorated index and the recipient.

6.7. Taste 011r! '!"t/chllcss

It could be objected that the argument of rhe preceding section is beside thepoint. Decorated objects please people because they confer aesthetic pleasure,and that is why the}' are desiI"J.ble, not because they are cognitivcly re."istant toanalysis, as I rune claimed. People just 'like' preu~ patterns. To this objectionI would answer th:u, first, not e\er}oody docs like prett), (complicated, anim­ated, etc.) patterns, and secondly that mere aesthetic 'liking' cannot cxplain thctypes of social relationships which are mediated by patterned artefacts.

J\esthetic pleasure is consummatory, an end in itself, and there is nothingempirically to show that the decorated objects with whieh the world aboundsare contemplated except in specific situational contexts in which their aestheticproperties are nc\cr thc sole focus of interest. i\1.elanesian big-men do not Silaround aesthctically contemplating lime-containers; on the contrary, they treatthem as mediating object's. Lime-containers are not self-suflicient sources ofdelight, but vehicles of personhood, to be owned, exchanged, and displayed.Aesthetic theories of ;lrt imply that any two recipients of :all artwork, if theyha\'e thc same aesthetic tastes (which would apply, ajimilJri, to two J\t{e1ancsianbig-men) will ha\'c [he same rcsponse to the aesthetic properties of any givenobject, But this is never the case in practice; so f:lr as these big-men ;lre con­cerned, the inherent quality prc.<>ent in an index (such as a lime-<:ontainer) ismooul:ated by the identity and status of the gourd's owner. The :aesthetic prop­erties of;) Iime-t.:ont'ainer are salient only to the extent that the)' mediate socialagency back and fonh within the social field.

Similarly, I would deny that the contemplation ofan oriental carpet that onedocs not own is the same kind of experience as the contemplation of :\ c.tr­pet that is one's ]lcrsonal possession. The possessor of a carpcl of complicatedOtiental design, as in the well-known Henry James story, sees in its coils animage of his own unfinished life. It is quite otherwise with sOlllebody else's car­pet, which is JUSt a snare for unrealiz;\blc desire, Aestheticians will no doubtbe horrified at the cxprc.<;<;ion of such sClllimcIlls, but :\llthropologiC'"J.lly, it isabundantly obvious that aesthetic responses are suhordimlte 10 responses stem­ming from the social identities :lnd diO'ercnccs Illediated by the index. Sincethe pure aesthctic response is a myth, it C':ll1not be im-oked to e.'{plain the \'cr\,

82 TII( eri/il/llt: oft/I( flll/n '3manifold types of attachment bct\\CCll persons :IIlU things. The ;\cslhctic

response alwa~s occurs within a social frame of some kind.Set:nndly, when it comes to the SIX:cifically Western (or, indeed, Oriental)

'aesthetic attitude' it is objects dc\ aid of precisely the kind of inrrictte surfacedecoration of the type I am intercslt.xJ. in mal have often excited the most extrav­agant aumir-.nion. Simple and undecorated forms arc the most beautiful accord­ing 10 refined canons of taSIC, and equally, to those whose religious <lttitudesincline them towards asceticism and withdrawa1. As E. Gombrich (198.t) hasshown, most committed aesthetes are far from keen on riotous decoration, whid1,JlOwcvcr, survives and prospers, cven in the face of aestheric condemnationfrom on high, because it is socially efficacious. I, personally, in my incarnationas a person of refined tastes, ,ldmire even to excess ultra-plain Shaker furniture,but anThropologically, I know that this furniture came into existence in a (om­munity which specifically, on theological groulH.ls, outlawed the kinds of medi­ation which decorated objects allow in h.'s5 puritanical social milieux. ShakerchaiTS \\ere made plain so that Shakers \\ould not be auached to chairs. or otherearthly things, but c..'l:c1usi\·ely to Lord Jc..'>w•. Kantian high-bourgeois aesth­etics cannot explain decoration became Kanrian aesthetics is agail/st de<:oration.

Most pt.'Ople arc not like Shakers, they prefer to load surfaces with decor.ationin order to dr.aw persons into worldlJ projects. An ornate Victorian chair is nothalf so beauriful as a Shaker one, but it commonicales much more strongly, notjust rhe idea of sitting comfortably, but also a host ofother domeslie and socialimplications of a this-worldly vnriety,

Gombrieh (H)8+: 17-32) has traced the epidemiology of pro- and anti­decoration sentiment in the history of European taste. I-'uritan movements,such as tIle re\'olmionary modernism articulated by Adolf Loos, despise sur­face ornamentation, while romantic hedonists sw.:h :IS the 19605 hippies whocustomized their Volkswagens with no\\'ers and S{:lrs, adore it. I am not goingto n..'Capirul:tre Gombrich's excellent history; instead I want to focus attentionon what I have identified as the essential property of surface decoration, itscogniti\Ce resistance, the fact that once one submits to the allure of the panern,one is liable 10 become hooked, or StucK, in it.

The fact that the word 'taek~" is rhe one selected (from a mngc of possibil­ities) by severe modernism to condemn the popul<lr taste for riotous ornamcntand other lapses of taste is interesting in itself. Let me quote from MaryDouglas's summary of S:lrtl'e'S account of 'viscosity' as an ignoble stare of

being (Dougl<ls 1966: ]8; Sartl'e t9.1-3: 696 If.):

The \'iseous is ~ st<ltc ha1f:'w<lv betll'cen solid ,md liquid. It is like a cross-section in aprocess of eh,mgc. il is unstab'le, hUT it docs nntllow. It is soft, yidd.ing: 31.1d e()Tl1prcs~­ible. There is nu gliding on it.. surf:K'(:. Its stickiness is a trap, it dlllb'S hke a 1cCl·h; ttattacks the hound3f\" berween m}sclf and if. Long l.'olumns falling olT my lingerssugg<:s1 Ill) own sub~tanee flo" ing into a pool of slickint.'SS ... to !Oueh slickincss is !Orisk diluting m~sdfinlo \-iscosil~ .. ,

Sartre is perturbed by rhe sub\'ersi\e effect of ,·iSt.1JSity on the body/worldboundary, but in the lighl of post-J\hussian understandings of gifts as adhc­si\-e components of persons, strung betwcen donors and recipients on loops ofviscous (if imaginary) substance, we need not be so squeamish. Physical/tactilcadhesiveness is perhaps genuinel~ disagreeable, bUI analogical, or cogniti,·e,adhesivcncs." is not, otherwise we would not be so willing to have artefact!;attached ro us and so positi\'ely responsive ro the adhesive qualities of surfacedecomtion, Most non-modcrnist, non-puritan civilizations value decorarive­ness and allot if :t specific role in the mcdiation of sodnl life, the creation ofattachment between persons and things.

6.8. The Ap(){ropaic Pal/em

The formula for ,he abstracl panern gi\en pre\'iously:

LLLlndex-A motif"""J ~ lndex-A.........:bok] ~ Inde.'I:-A whole]~Recipient-P.

refcrs to the cs.~ntiall~ hier.archical nature of the index, thc fact that it is com­posed of parts combined into a whole (this applies to all indexes, but is par­ticularly salient in this contexl). My argument is that the deconstruction ofthis complex of hierarchicJ.1 relationships endows Ihe decorated objecl with acertain rype of agency, which is the reciprOC:ll of the agcncy exercised by therecipient in (attempting) to perceive it; his action is subjectively experienced asa passion, a pleasurable frustration. However, other types of agency m,l)' bepresem as well. In particular, the agency of the artist may be abducted frOI11 thepattern. Thus ,11e formula can easily be expanded to:

[[[[t\rtist-A] ----) Index-A lllotip··.. ] ----) Index_Al"'n/-t...I<] ----) lndex-A whole]~ Recipient-I'.

This agem:y can be agonistic or defensiyc as well as beneficial. Irurn no\\ toa series ofexamples in which the exercise of social agency turns on the employ­ment of patterns as mcdiators of aCli"e!y hostile, or defensive, intent. It mightseem pamdoxical that patterns, \\hich bind persons to things, should be poten­tial weapons in situations of connier. A moment's renection is sufficient torealize, first of all, that relations of conniet and strugglc arc jusr as 'social' asrelations of solidarity, and secondly, that where\'Cr one finds conflict there onefinds abundant deployment of all kinds of deeorativc art. Much of this ;Irt isof the \'ariety known as 'apotropaie'. Apotropaic art', which protccts ,In :lgent(whom wc will take to be the artist, for the present) against the recipient (usu­a11y the enemy in demonic rathcr than human form), is a prime instance ofartistic agcncy, and hence a topic ofcentral concern in the anthropology of art.

The apotropaic usc of pancrns is as prorective devices, defensi,'C screens orobstadcs impeding passage. This 'apoLropaie' usc of pauems seems paradoxical

L-'----- J ......I

84 The Critil/II" IIflhe index The Criliq/lc III/hc index 85

PIG. 6.811 _ Celtic kno!work: apotl'opaic

p;1ttern

FIG. 6.911. Koll/ln threshold <i",~ign: the pattern~s topologic~l snare

in that the placing of patterns to keep demons at bay seems contrary to the useof patterns in other contexts as a means ofhringing about attachment betweenpeople and artefacts. If patterns attract, wouldn't they also attract, rather thanrepel demons? But the paradox is apparent rather than real in that the apotfopaicuse of patterns depends on adhesiveness just as the usc of patterns to attract peo­ple to things. Apotropaic patterns arc demon-traps, in effect, demonic fly-paper,in which demons become hopelessly stuck, and are thus rendered hatmless.

Take the Celtic knotwork pattern in Fig. 6.8/1. Knotwork like this wasregarded as protective in that any evil spirit would be so fascinated by theentwined braids as to suffer from a paralysis of the will. Losing interest inwhatever malevolent plan it had entertained previously, the demon wouldbecome stuck in the coils of the pattern and the object, person, or place pro­tected by it would be saved. Not just intricate pattern, but sheer multiplicirycan hav~ this effect; I am told that e\'en recently in Italy, peasants would hanga little bag of grain next to the bedstead, so that the Devil, approaching thesleeper in the bed, would be obliged to count the grains in rhe bag and wouldbe thus diverted from inllicting harm. The interminableness of large numbersand complicated patterns work in the same way; but patterns arc more inter­esting and certainly more artistic.

6.9. KolamA suitable example is provided by the kind of auspicious threshold designsknown in Tamilnad (South India) as ko/am, and by other names in other partsof southern and eastern India where similar threshold designs are also drawn(Fig. 6.9/T). In Tamilnad, kolam have two protective functions. F~rst of all theyare associated with the protective, fertile, and auspicious cobra deity (lIaKa) andsecondly they have an apotropaie function, repelling or ensnaring demons.

Kolam are made by women, who dmw the design free-hand in lime, rice-powder,or some other white powder which they trickle through their fingers. Thedesigns are made at dawn, especially during the time of year when there arcmany demons and spirits about, and arc rubbed out by the passage of feetas the day wears on, to be renewed the next day (Layard 193T 12t, citingJ. Dubois, and S. M. N. Sastri). Fig. 6.9/1 shows a typical ko/am design. Ko/amare sinuous, symmetrical figures which are usual1y very difficult to 'read' in thesense that it is difficult to see how the design has been constructed. They playtopological games with the spectator and are thus akin to mazes. Let us takeFig. 6.9/1 apart and see what kind of cognitive teasing is going on here.

On first inspection, it seems to consist of a single line, pursuing a complex,sinuous path between the rows and columns of dots. However, this is an illu­sion in that this ko/am is actually composed of four continuous loops of asym­metric configuration, superimposed on one another while being rotated in90 degree steps. This to/am is, so to speak, the visual equi\'alent of a canon infour parts, in which each voice sings the same notes but not in phase; that is,after the lapse ofone bar as in 'Frere Jacques'. The identical components (indi­vidual loops) of this arc out of phase, not in time, but in space, being displacedby 90 degree rotations (Fig. 6.9/2).

What is most interesting about this ko/am is that even if one 'knows'(intellectually) that this design is made up of four separate, but identical loopsdilTerently orientated, it is extremely frustrating to attempt to abstract incli­vidual loops from the ovcral1 design. And though I find I can, with some effort,do this for one loop at a time, I cannot do it fur two loops simultaneously, stillless three, or all four. Where simple geometric figures are concerned, 'seeing'(a triangle, say) is tantamount to mentally intending or projecting the con­struction of the figure, line by line. But this is impossible to do with complexfigures like this ko/am. Here seeing the figure is quite distinct from being able

86 The Criliqu~ IIfih( II/do Hi! Critique of/IIC }1/(!c.1 87

FIG. 6.9/::. The ronstilutin~

element or the h/llm

to memally intend the process urits construction. Yet w(; 'write in' the fact thatit is possihle 10 construct the figure, bewusc here it is, it has !:){.'cn made by some­one, and we might indeed have been lucky enough to hare watched, entranced,the deft mo\ements orthe \\oman who made it. So we end with a series of para­doxes. We want to see this figure as one continuous line, but we know it is fourseparate loops, which, however, we cannot separ.uely abstract from the overalldesign. We know too that this design belongs to the Teal world, and was made,apparenrly effortlessly, by this real woman; bUI we ClonOI retrace fully the pro-­cess whereby the design came into the world, b) the agency of this woman,because we cannot reconstruct her skilled movements (and the imenrions guid­ing them) from the design vvhich has resulted from them.

1 attribute the cognitive stickiness of patterns to this blockage in the cognit­ive process of n:cunstTucting the intentionality embodied in artefacts. Clearly,no demon will readily cross a threshold set with sophisticated topologicalsnares such as this; the evil one will have to pause lHld ponder the problem, andthe impetus of its malevolence will be neutrali".cd. Perhaps it is no coincidencethat Tamilnad, the state where women play these mathematical games, is alsothe state from which most of India's world-renowned mathematicians andcomptller wi:.·.ards originate.

6.10. Kolam, Tattoo, ami the Cretan Maze

In the earlier part of this cemury, designs frequently identical to 1:01am had analternative usc as tattoo designs, defending not the threshold of the house, butthe skin, the threshold of the body. And [his provides a link with the nexr s[ageof my argument, which has to do with mazes. Fig. 6.to/J raken from Layard1937, shows a South Indian talloo design published at the beginning of thiscentury, and for comparison, a 'map' of the labyrinth at Knossos (the Olle con­taining the ;'-.·Iinotaur) as depicted on a Cretan coin of the seventh ccnTUry Be.Lcgendarily, this labyrinth was designed by thc master-<:raftsman Dacdalus,and was modelled on anothcr mazc, the onc which Icd LO the Underworld. Asimilar reprcscmation of the Cretan labyrinth is mentioned in book 6 of the

•

FIG. 6. l0ll, Compariwn of Ihrcshold,l~tlOO, ;un! ~oill designs from La\'ard193i: fig. 36: '(Ill threshold design,;"·,,ltllll, "the fUrl"; (h) lattoo pallenl,"fort": (c) clIm elltionali,.<XI Crelanlabyrimh design appearing-Oil lhecoins of KnQSsos (for oompariJ.oll)'

•Aeneid, as being carved on the Curnne:m gates, where it attracts the lltrenrionof Aeneas' men, prior to theil" leader's descent into lhe Underworld to consultthe Cumae:lll Sibyl, who predicts that one day Rome will rule the world.Curiously, :In exactly identic-dl map of the lab\"tinrh is to be found Co1r\'cd nc.\t~o the entry into the grcat passage-grave at New Grange, in Ireland, which ismdeed also a twisting tunnel leading, prcsumablv, l'O the world of the dead.There arc also representations or gods at the celll~c of labyrinths in the art ofancient Egypt.

Why do I mention these facts, which so excited the anthropologist Lanrdand the classical scholars of his period? La}ard is now a forgotten figu;e inanthropology, :md his interest in matters such as Ihese ensured the demise ofhis reputation in an era dominated bv Ihe fllnctiomllism of Radclirre-Brownand ~\hlino\llski. Layard was lhe I:tst of the diffusionisrs of a psychological per­suaSIOn, who (lre\\ their theoretical framework from Rivers, and now:ubvs theonl~' Layard wor~s in print are thosc he wrote blcr on as a Jungian :ll;alyst,whIch are 10 be found, I am told, in 'Occult' bookstores. Onc can sec whvLayard abandoned academic 'lnthropology and became a proponent ofJ ungia~'arch~t~ p~' in the lighl of the facls just mentioncd. It is indeed astonishingthat ldentl(."31 maze-paUcrns should be une-.anhed in Soulh India, the ancient.\ Ledilcrrancan, E.g~·pt, and lhe bogs or Ireland, especial I} so in that, as \ve v\ ill

1lte Critique of/he fl1fkx

see, J ,a)'al'(l found similar ideas linking 'm,lze' designs amI entry into the landof the dead .llso in the New Hcbrides (Vanuatu) in the midst of the Pacificocean. I feci sympathy for L'Iyard in that he was in a sense 'bounced' into pur­suing a deeply untrendy dilTusionist approach (for the time) b~' ,he sheereohereIH.:e of the dat'l at his disposal, for which there docs indeed appear noob\ iOlts explan'llion mher than a Riven;ian or Jungian one.

I pcr!iOnally do not see why one should dismiss diITusionisr expbnations OUtof hand (cc. Lth i-SmlUss 19(3). But it is nor the diffusionist hypothesis thai I\Iam tel discuss here. Enn if the 'Cretan maze' idea has spread across the globefrom some specific place oforigin, one would still ha\'e to explain the enhancedreceptivity uf vcry diverse civilizations to this idea, on a bener basis than theprcscn"ation ofancienl racial memories, as Jung proposes. One can afford 10 beagnostic aboul hO\1 exactly the labyrinth idea came to exist in so many differ­ent places with apparently the same associations, ha\'ing to do with the pa.~sage

between this Ilorid and the land of the dead. "l'he problem I am concerned wilhis to unden.tand the cognith'e significance of thi.~ mazc-panern, which couldequally well explain wh} it ha.~ tranl1cd so far and wide and/or why it has beenill\entcd independently in so many muruall~ faraway pla<."t.'S.

The Cretan labyrinth is not, technically, a 'maze' in that it has no branchingpaSSlgeways and so poses no navigational problems. Theseus would not reallyha\"e n<.'Cdt.'d Ariadne's thccoad to find his way inro or out of the Iabyrilllh if itwere built as sho\\ n in Fig. 6.1011. It is a 'meander' pattern, an exercise infinding the longest possible pathway between two spatially quite adjacentpoints. It evokes the idea of a na\·igational problem without actually offeringone, rather, thc problem is one of distance and the continuum. One (:ould s(''Cthe Creran maze as a spatiaJ version of Zeno's paradox<.'S of time and motion.Zeno argued, for insrance, that an arrow in its flight must successi\ely reach,and pass, an infinite number of points, that is, the half-way point, the three­quarters point, Ihe6/8 point, the 12116 point, the 2-+/32 point, etc. etc. and thaisince this series of fe-Jctions is infinite, the arrow will never reach its destina­tion. The Cremn maze is built lip by connecting points serially (Fig. 6.1012) soas to creare a meander pattern which threatens to indefinitely extend the num­ber of twistings and turnings in the path which must be traversed between thecntn'-dool' and the centre. In the Roman world, meander patterns (and genuinemaz~-pallerns) were .1 popular device for mosaic floor-decoration in palaces,where their capacity to expand confined spaces indefinitely Illay indeed havebecn al'l.'hiteeLOllically highly effective, as well as decorative (see rig. 0.10/3,showing the great maze at PU!:I, in Croatia). Such mazes exemplify .mother \\,.IYin whidl patterns call present cognitive obstacles. One knows there is a waythrough the maze; one may cven know that the m:lze is created by the simpleapplication of an iterative rule in connecting up lines and points, as is demon­strated in Fig. 6.1O{2, but onc eann()[, all the same, see one's way through themaze except I'cry laboriously by tracing out its winding course.

'Ole Crilif/llc !J.l'the fllil,'x

:.J~.'1 . -

89

FIG. 6.lof~. The lienalconnection of poims inme fonnation of {heCreW! lah~'rimh drsign

FIG. 6. I0/3. The RomanmaZe al Pula

1)0 The Crililjue oldie Inda The en/IIjUi' /lIthe !mk.>; 91

Th. child'en are:

\\\ f."",re cycl••

II

FIG. (j,ITII. Diagram traee;,d by IheAmbrym natives to explain their kinship'ystem. (The long lines stand for men,and Ihe short for women; the arrowsseparate Ihe 'lines' wilhiJl the samebilateral group.) Caption and diagramfrom Lcvi-Strau:;,s, 'lhe F!~menl{iry

SlrU(WreJ ofF(imhip. Source: Lhi­Strauss 196<): fig, 5

FIG, 6.11/2. I.e-,i-Strauss's diagram toexplain Aranda kinship in Tile SlI1,ageA·lind. Sourre: Levi-Strauss I')j2

AmanAl(from ""'t'ilineal moiety 1)

ma,rift.a woman C2

Boyo A2 (COtl'inuou. I"",)wtIo many women OJ (b,,,,,",,line. out... c~cle)

G~!. A2 (b'oken find"".,ci«cleJ ","rried by men01 (OOI"Iti....""'. Ii....)

Motriline.ol moiety 1

~-0--~ ,7'" A '''\.A

I ,'~... \I ~ c, \, '\" \ \~ M.t,ilineal ~~ moiety 2 ~

\ ' I I\ ,\ ....0 )f /

\ ----" /, G 'Ca, B ",/, ~....... _.....

Social .t'u<:lu endma......g" ",I ofthe Ar.nda type

6.1 I. Smui-DrallJings ufA1alakula alld lhe Land ufthe Dead

This lTlay help explain why mazes are associated with the passage betweenthe worlds of the living: and the dead. These worlds are close together (de:lthis everywhere) yet (Ir apart, separated by an impassable frontier. Aeneas' com­panions were prevented from entering the Cumaean gates by the fascinaTionexerted over them by the maze design carved on the portal; in this respect theywere akin to the dcmons waylaid by the ko/am protecting thc house, discussedearlier. But the connection betwccn koltlln and maze designs is closer than this,in that Indian (kollllll-based) tattoo designs were specifically connected withentry into the world of thc dead. Thus the tattoo design shown in Fig, 6. I 0/1,

besidcs being formally identical TO a ko/am, is unc of the class oflndian tattoodesigns, occurring in many regions of the peninsular, which arc idcntifiedas the 'fort' (i.e. City) of the land of the dead. Throughout India, especiallyamong lower castes and tribes, it was considered a neccssity for women, andsometimes men, to be tattooed, in order to avoid punishment in the land of thedead. Thc 'fort' design was considered to help the dead person (whu retainedher tattoos pOil mortem) to find their way to the bnd of the dead and be safdyreunited with deceased kin. So the 'fort' is a map. i\t the same time, it is alsoa puzzle, in that these same people believe that Yama, the God of death, andhis demons, will devour the untattooed, bur will not harm the tattooed womanbecause they cannot solve the 'puzzle' that her tattoos present, that is to say,the tattoo is aporropaieally protective of the body after death (Uen n.d. [I994J).SO the kolalll and the 'maze' tattoo both work in the same way-both stand forthe idea of (demonic) attention being drawn into the pattern, being tantalizedby it, while being thus rebuffed and rendered impotent.

The idea of patterns as 'obstacles', that is to S;lY, sticking-points, surfaces alsoin Malakula, the l\ew Hcbridean society investigated by Layard and his oldercontemporary l3ernard Deacon, who died while he W:lS in the field (bothDeacon and Layard were pupils of Rivers). Malakula arc perhaps unique inhaving contributed to the literature of anthropology not just as ethnographicsubjects, but as originators of a metllOd of diagrammatic representation in kin­s11ip theory. The diagram drawn by a Malakulan man for Deacon (using short:llld long sticks, and arcs traced on the sand) showing the operation of Thelv1alakulan three- (or six-) class alliancc system is plainly ancestral to subse­quent anthropologists' efforts to create diagrams of thc same kinds of institu~

tions in ,'\ustralia (Deacon I934: Fig. 6.I1!I; Levi-Strauss I96(j: Fig. 6.11/2).The reason why the .Malakulans were so diagram-minded is that they practisedpattern-construction as an art form and as part of the typical Vanuatan malecompetition to demonstratc knowledge and mastery. Deacon collected morethan {imv-five designs which were published after his death (Fig. 6.tt/3). Thesedesigns ,:'ere drawn on the beach by men, and would of course he washed away

is to outwit the demon by succeeding. We arc a long way, here, from the idea thatpallcrns :lppeal \() the eyes or give ,H,:sthetil,; pleJsurc. I do not think that the.\13131;uI30s thought of these patterns as independent \,jsual ubjct..-ts at all, but asperformances, like dances, in which men could rcn:al their capability. :\'lelanesianaesthetics is about effielc)', the capacity fO accomplish Tasks, not 'bcaut~"'.

6.12. Drawing (/ud Dal1cing

At this point one may draw attemion to another source of insight in Layard'swork, which was also unerly lorcign to the mind-set of his anthropological con­temporaries. Layard (1936) discusses the affinity between the choreography of.\l3lal<ulao dance and rhe slyle of their graphic arlo The ghost's task of com­pleting [he maze-likc figurc on the sand, is complemcOIcd by the performanceof ceremonies conncctcd with the inducTion of ncophytes into thc men's cultin thc course of which a single dancer (called 'the 11awk') has to thread his waythrough the ranks of the main body ofdan<:crs very much indeed as ifhe werenegotiating a maze (Figs. 6.12/1, 6.1212 from Layard 1936). Layard links this

-liII,,

TM CNl,qll~ 'yfltt' l"do 95

with the st)-C\lIcd 'Trojan Games' described in the Iliad. equcsTrian militaryexercises, witnessed hy the Greeks, in which ll1ou!1Ted youths executed a seriesof circuitous manreunc." supposedly tracing out the path 01':1 labyrinth, Butonce again it is nolthe diffusionist thLosis which concerns me, bur rhe nalure ofthe cognitivc linkages being suggesled hcrc_ It is surely useful to consider thcact of drawing as akin to dancing, :md the design as a kind of frozen residuc leftby this manual ballet. Indeed, jusl this .malogy seems to h.I\'C suggested itself10 .\kTleau-Pomy. after \\;tnessing 3 close-up, and slowcd-down, cincmato­graphic record of Matissc's hand and brush eng-J.ged in the act of painting.

This dancer-like nimbleness of the hand (and indeed the analogy with thehawk. The mosT precise of birds in its movemenls) emerges as a cruci:ll elementin thc e\'alu3tion of plasTic (sculptur-,d) art in the Trobriands, for instance (cf.Gell 1992h). But, enn more, it indicates thc synerg} betwL'Cn art forms andmodalilies of c:~pn.........<;ion which cOlwenrional aL~lhetics tries to deal with separ­,Ilely, because they givc pleasure 10 separate senses. the eye (visual.Irt), the ear(music), or lhe "inet'ic sense (dance). If we can see visual parterns as frozentraccs of dances, so \\C can also see dances as being half-way to becomingmusic, which indL'Cd nonnaUy accompanies them. What unites drawing, music,and dancing is a certain cognitive indceiphenlbility manifesTed in performance(see above, Sect. 5.2). Thus, to revert to the comparison made earlicr betweenour 1:01am design and a four-part (music-J.I) canon, we may observe thaT a fuur­pan C:l.110n reveals its strUCture (because il is L'3.."y to hear Ihe four succL'SSivecntries of the theme) blH also conceals it, in that it is ncar-impossible to hear.111 foul' parts simultaneously. So :lIso the kolam rcvcals itself as constructed outoffour superimposed figures, but just how, we cannot be certain. Drawing andmusic and dance tantalize our C:l.pacity to deal with wholes and parts. continuit}and disconrinuity. synchron)' and succession. This analogy also reminds us ofthe fact that much art consists of \'irtuosic perfollnances, and that although theperformances involved in most \'isual art Take place, so to speak, 'ofT-slage',nunc the less a painting by Rembrandt is a pcrformalll..-c by Rembr-J.ndt, and isto be understood only as such, just as if it were:l peTfonnance by one ofroday'sdancers or musicians, alive and on-stage.

1-,­,,­,

II\ ""'---,-----... "'l, " .... _--_ .. '

.,.. -------- ....: / ....- -i- i-i-i-,',f: ,.----------" :"',. ~ • t i ,II --------- .... I

" • • t • .',,I JJ ,

I :'~ • i • • :1,"---------... II". I • t t;,\ .. _" I' ...-------- -"'"--- - ---

94

oS-...--~, ,

-' II , ,____ ', I I

,- '.\' ,, I I I ,, .. I I

I _... I JI ,.•.~__:==- -=. ==:C"_-_:::.:-,.'/'~~!l_t_!.. !2-'-~~.'.··'I' ·· .. ··~I······· I ,t r·_-_£_-__·_.!,,_!.._·_!._!./l l1'-:··············· ...·········· .. ··.1,I ..~_._._~_._. ~_._~_._:I, ., ·.", :~~_._~~~---~_._~' III I • ........................... " JII .. ~~_._'!. <tJ._!..~_~~:. ,I :-:t':•••••• .........- I I

I II, ;'r'.:-.:-.:""••'";..:.7.::.:::: ::.. :::-.. II, .. !,. ~~_'!..! ~_._~_._~: I,

................................ ~ I 1II !!.!....!_._!..~_._'!..~_-:.' ,I I I

"!.~~_·_!._·_·_.!_'!..!',)I ,I I : ••; •••• ;, ; ;.; I'\ ' ..'.- - - - - - - - - - - - ..~ I, .. _----------- "",'.._------------

Ft<j. 6. 1~II. I'erlormanec urt mmpicmeniing[(T1Iphic an: ,he Mal:,l.:uLm Ilan<.:er's palh. 'The

dOls repTCSCIlt Ihe main bod}' of dancers" Ihe:!.Trows reprcsentin! th{'" dircction in "hichthe\" arC fadng. Th{'" brulen lines represcnl Iheparh tnIce,] b~' the I [awls. The doued linerepresent.' lhe "j'!:lwks' return track'" (L:ly~rJ

'937: fig. 24)

FIG. 6.r2/2. 'The Hawk flying from plaecto place' (LayarJ 1936: fig. 2..)

7

The Distributed Person

7.1. Mimesis ami Sorcery

I now rum ro the aOlhropological theory IIf rcpn.:scOI:ltiunal art, that i~ to say,the production, circu!:Hion, and practical use of indexes thai h,l\"e reicHm pro­rotypcs, something that geometric pattern!> lack. !\'losl Western art is of thistype, and mos( of an-theory is about representation, in one way or another. Ishall try to avoid, as far as I>ossible, discussing Western arl, tempting thoughit is to do so. Instead, this part of the :lrbTUmCm pivots around the cbssicallv'anthropological' themes of sorcery and image-worship. .

The basic thesis of this work, to recapitulate, is that works of art, images,icons, and the like have to be treated, in the context of an alllhrrlpII1IJg/m!theory, as person-like; that is, !>{lur<:es of, and targets fOl', s{)<:ial agen<:y. [n thiscontext, image-worship has a central place, since nowhere are images moreoh\'iolL"ly treated as human persons than in the context of worship and cere­monies. In this chapler, I present a general theory of idolatry, a practice I byno means regard as more misguided than any other religious obsen-ance; and13m, I hope, capable of sust3ining 3 sympathetic attitude toward<; religion ingeneral, despite not being a religious person myself. 'ldol3rry' has had a badpres." since the risc to world dumination of Christianity and Islam, whichh;lxe both inherited the anti-imagistic strain of biblical Judaism. Christianity,encuml>cred with it.<; Gr.K'co-Roman inheritance, has h.ld to struggle moreat..'ti\·e1r with recrudescences of dejiu/o 'pagan' idolatr}, and has experiencedcataclysmic episodes of iconoclasm; Islam has morc consistently resisted thelure of image-<lirecred forms of worship. ProtestanTism, the most dynamicbmnch of Christianity in the past few centuries, has hardly l)(:en less puritan­ical about idolatry than Islam, and the consciem::es of the Clfholics halT beenoften pricked on this account as well, so that the net rcsult is roday that 'idol­atry' is a pejorative word, which anthropologists, especially ones who claimto empathize with the religious sentiments of others, arc nor supposed to use.8L1l rather than resort to some vague or misleading circumlocution I prefer tocall the practice of worshipping images hy its true namc, deeming it betrcr toexplain idolatry, rather than rcchristen it-by showing th:lt it cmanates, notfrom slUpidity or superstition, but from the s~me fund of sympathy whieh:llluws us [0 underSland lhe human, non-artefactual, 'other' as a eoprescnlbeing, endowed \\'ilh awareness, intentions, and passions akin to our own.

Tit,. Dislnbllird Pers/!/t

Granrcd th:1t ido!alr)' may he so called, I still ha\'c to explain why a bookostensibly about 'art' has to dC"Ole so many pages to a lOpic which appears to

belong to the studY of religion rather than acsthetics. Those who object to mvanti-ae....thetics sta~ce will reg-drd as irrele\-ant the 'religious function' of image~ther prefer to sec in a ttJtally non-religious frame of reference, from the stand­point of aesthetic COnTemplation. An image viewed as a source of religiouspowcr, sakation, exaltation, is not appreciated for iL<; 'beauty', but for quite dif­ferent reasons, the proponents of 'aesthetic experiences' would argue. I reg-Jrdthis as fallaciOlL<; on two grounds. First of all, I cannot lell belwttn religiousand aesthetic exaltation; arl-Imers, it seems to me, actually do worship imagesin most of the rele\...nt senses, and explain away their de jacto idolatr} by ration­alizing it as aesthetic awe. Thus, to write about art al all is, in faCT, to \\ riteabout either religion, or the substitute for religion which those who have aban­doned the outward forms of received religions content themselvcs with. TheProtestant-Puritan heritage combined with ;I special form of art-theoreticalcasuistry have established a special form of b:ld faith about the 'power ofimages' in the contemporary \Vestern world, as Freedherg (1989) has pursuas­ively argued (sec below). We have neutralized OllT idols by reclassifying themas art; but we perform obeisances before them every bit as deep as those of themost committed idolater before his wooden god-I specifically include myselfin this description. Secondly, from the anthropological point ohiew, we havcto recognize That Ihe 'aestbetic attitude' is a specific historical produci ofthe religious crisis of the Enlightenment and the risc of Western science, andthat it has no applicability to civilizations which ha\·e not inrernalized theEnlightenment as we ha\·e. In India (which figures largely in the ensuing dis­cussion) idolatry nourishes as a form of rc1igiosil~, and nobody in thcir rightminds would try ro dri\e a \\edge between the lx:autiful form and religiousfunction of venerated idol..... In India aestheTics, as in the ancienr world, is sub­suml..-d under the philosophy of religion, that is, moral philosophy. as a matterof wursc. Consequently, it is only from a \ery parochial (blinkered) Weslcrnpos[-F:nlighrenment point of \·iew [hat the separJtion bctween The beautifuland the holy, between religious npcricnce and aesthetic experience, ;Irises.Since this is so, the anthropologisl writing ahOul art inevitably contributesto the anthropolog} of religion, becausc the religious is-in some contexts,though not aU-prior to the artistic.

Idols come in m;lllY varicties, but it is con\'emional to distinguish twO polartypes; (i) purely 'aniconic' idols, such as the bailulill (bla<:k meteoritic stOIlCS)worshipped in .1l1cienl: Greece, versus (ii) 'iconic' idols, that is, indexes physic­ally resembling ;1 prototype, usually .1 human being, according to the formula[[Prototype-A --+ [Artis1-A]] -----l< lndcx-P, where the prototype is the god,wbose 'likencss' is mediated by the artist. llo\\"c\"cr, before starting, it may be aswell to darif~ our reasons for not pa) ing \'ery much auen1ion to this distim:tion.:\11 idols, 1 think, are 'iconic'-induding the so-ell lied aniconic onc.<;-whether

The Dislribuud Paslm

or not they louk like some filmiliar object, such as a human body. An aniconicidol is a 'n:alistic' representation of a god who either has: no form (an}"\ here),or has an 'arbitrary' form, in the particular 'body' he inhabits for the purposesof being worshipped by his mort'<ll de\ Olel."S, here belm\. A meteoritic stoneis not aver), wry, com"emionalized or distorted 'portrait' of a god, who, e1sc­\\here, looks like a human being. One need not imagine that worshippers ofstones would 'prefer' to worship more realistic portTa)als of their gods, buthave to make do with unshaped stones for lack ofany local stone-C'J.rvers of thenecessary ability. For such worshippers, the meteoritic stone is an index of thegod's spatio-Iemporal presence (and origins, in that the stone fell to eanh fromheaven). The stone is a 'representative' of the god, like an ambassadur, ralherthan a visual icon. Just as an ambassador call represent his country in Moscowone year, and the following year in Washington, the stone represents the godat whatever spatial coordinates it happens to be; and it is movable. One such'black stone' for instance, \\orshipped in ArJ.bi;1 as Cybele, was com·eyed withgreat ceremony to Rome, where it \\as installed inro a Cl\'ity within the idolof Magna !"Iatcr Idea, the goddcss of members of the Imperial elite who fav­oured an 'intcrnationalist' religious outlook (DumcziI1980; on thc significanceof such transactions, see belo\\). The ideas of 'repfL'SCnting' (like a picture)and 'representing' (like an ambassador) are distinct, but none the less linked.An ambassador is a spatio-temporally detached frugment of his nation, whotra,-e!s abroad and with whom foreigners can speak, 'as if' they were speakingto his national government. Allhough ambassadors are real persons, they arealso 'fictions', like pictures, and their embassies are fictional mini-states withinthe state; JUSt as pictures show LIS landscapes and pcrsonages who arc 'not reallythere'. Although the Chinese ambassador in London docs not look like China,or the Chinese government or people, he does JUJu If) he visible, and he docs·lJisibly repreSUlf China on official occasions. He does not look like China, butin London, China looks likc him.

One could not contrast a 're:alistic' iconic ambassador 10 an 'unrealistic', ani­conic amba.ss.1dor, and no morc, perhaps, should onc contrast 're:alistic' idols,to unrealistic, aniconic idols. Whatcver the idol looks like, that, in context, iswhat the god looks like, so all idols are equally realistic, because the idol-formis thc visual form of the god madc present in the idoL The contrast is notbetween idols \I hich resemble human beings to a greatcr or lesser dl.'gree. burbetween gods who in idol-form (visually) rl...-.;cmblc human beings to a grcafCror lesser degfl.>C. Idols, in other words, arc not depictions, not portraits, bur(anefaetual) bodies. The formula cited above then r[Prototype-A ---7 [Artist-A1.1-----1 Index-P docs not necessarily imply 'realistic' (i.e. anthropomorphic)depiction of the pmtotype by the artist in the form of the image. When sucha formula is applied to portraiture, this impliGltitJn exists, bcGllISC there isa living model for the portrait, and fidelity to lhis living model is the soci;llobjecti\e of ponrairure. But in the context of idolatr~, the agency in\ol\·ed is

Tht Distributed Persoll

religious. Where an idol is an artefact (rather than a natural object, such as ameteoritic stone), the nature ofagency excrted by the prototype is to ClLlS(: theartist to produce a religiously stipulated image according to the conventions forIt"Uch images, which may be iconic/anthropomorphic or abstract and anieonic.In either case, the artist has ro produce a 'faithful' 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, rhe idol is not a 'depiction' of rhegod, but the body of the god in artefact-form is :111 \'cry well, bur I accept thatany such assertion constitutes a paradox, and I must labour hard to dispel thepuzzlement it must inevit:lbly produce in the mind of any reasonablc person.The mystery of the animation of idols, their genuine, if peculinr, personhood,has to be approached, like any difficult problem, via a series of incrementalsteps, not rn mcntion detours through some unfamiliar territory. Rather thandeal with ido!;nry in its most elaborated fonn, I shall introduce the subjl.'Ct bypresenting an analysis of mit sorcery (mvoutemen/)-the practice of inflictinghann on others \'ia their images. This is a particularly pertinent case ofagencymediated \Cia representational indexes, and it has the ad vantage of removing thediscussion, temporarily, from the sphere of 'religion proper', which ill\'oh'csthe always problematic quesrion of other-worldly entities such as gods. Voltsorcery takes place in this world and without the necessity for other-worldlydi\'ine or diabolic intervention, though it may be soLight. As we will be able toobserw, there is actually a smooth transition hctwl.'Cn this-worldly volt sorceryand religious image-worship orientated towards other-worldly beings. Thiswill be explained in due course.

7.2. The /vlimetic Facully

Frazer, rather than Tylor, is the anccstral figure who presides over this dis­cussion. Frazer, it will be rec-.llkd, distinguished (\10 basic modes of'magical'action; (i) contagious magic, the magic of contact, in which influence passesfrom one object to another, and (ii) sympathetic magic, which depends onshared properties, that is, if object A shares properties with object B, A hasinfluence over D or "icc \-ersa. The Frazerian idea of 'imitatj,'e' sympatheticmagic has had an enonnous influence on aesthetics and the philosophy of artin the course of this century. Vial3enjamin and l\Jorno, this influence seemsset to continue; Taussig (1993) in a most exciting recent book, whose primaryinspiration comes from Denjamill, has this to say about Frazer:

1am particularly taken by rFrazer'sj proposition that the principle underlying the imi­tative component of sympathetic magic is [hat 'the magician infers That he can produceany effect he desires merely by imitaling it'. (52) Lea,'ing aside for lhe moment thethom~ issue of how and wilh \\ hat success Frazer could put himself into the head ofone of lhese magicians, and to \\ hat degree either the a(:curaq or usefulness of his

100 The Di,mbulCd Person The Di$/rilJllted Pasoll 101

proposition dcpcnJ~ on such a move, r want to Jwdl un Ihis I/o/ioll of/III' COP}', ill 11111).:­

ital flrat/ia, ailet/illg Ih.. IIri;:inul /I) .fIlCh IJ dcgru thul til,' reprtstllllltioll slw,ts ill oruCiflllres lilt' prOpUlit$ of/he r;:pri'St'lIwl. To me this is a disturbing notion, foreign andfascinating 1101 bClllUSC if so t1al;r.mtly conrradicts the world about me but T:uhcr, thaT0111,:C posiu:d, I suspect if nOI its pr.:seJlce, then intimations Ihereof, in the sl'r.lllgclyfamiliar commonplace :lIld unconst:ious h:lbits of represemation in tht: world ahout me.

(Taussig Il)()j: ·17-8)

Following Ucnjamin, Taussig bases his analysis of mimesis in the colonialmilieu (of the Cuna Indians) on a supposed 'mimetic facult)". Hcnjamin thoughtthat the mimetic facult\" \\ hieh in modern times has r<:suiLed in a world filled too,·crflowing \\ith imag~ and simulacrJ., so that norhing seems n:al :10)' more, hadiL<; origin in a primith'c compulsion to imitate, and thus g"J.in access, to the work!:

(.\Ian'sJ gift of SI.:cing- rcso:mbl:mces is nothing other Ihan :l rudimelll of the powerfull:ompulsion in former times 10 bnXlme :lnd behare likc somcrhing dsc. Pc:rh:lp!> thereis none of hi... higher functions in which his mimetic [:leulry docs not pl:l~':l dl'Cisi\'crole. (1933; cited in T:lus.."ig 199.): 19)

Taussig's book bears witness to the productivity of Frazer's ilk-a mediatedthrough Rcnjamin's surrealist imagination, bur it is fair to say thaI the 'mimeticf.1culty' is only rathl'f "agucly delineated. The fact that so much human bcha,'iouris imitath'e does nOl necessarily imply the existence of a 'faculry' inheritl't.! frumthe distant past. Almost allieamed behaviour could be dl'SCribed as imitative, inthat it is based on the imitation of an inl'ernalized model. J\limesis, narrowlydefincd, in\'oh'es the actual production ofimagcs (indexes) whose salient proll­eny is prototype, ,'ia resemblance to the original, and within this category artl.'­fa(,.~, haloing visual resemblance to thc originals, em be accorded a separate statuS.

What Frnzer ne'·er explained is why the mutual resemblance of the imageami the uriginal should be a conduit for mutual influcnce or agency. Ill' atnill­llted it to a mistaken hypothesis, akin to a scientific theory, hut grounded inerror. The trouble is thaI if the practitioners of sympathetic magic could haveseen thcir praetia:s a~ Frazer saw them, they would never have enb'ilgccl in themin the first phlce. Uy abstracting a generalizable 'principle' from the inchoateworld of practice, Frazer guaranteed his eventual misunderstanding of the (htt~l

he had at his disposal. Taussig, in his Benjaminesque reanalysis, ~lrgueS thatthe basis of sympaThetic magic is not a tragic misunderstanding of The mtllreof physic:'!l c:;lIs;lliTY, hut a consequence of epistcmic awareness iTSelf. To see(or to know) is to be sensllollsly filled with that which is perceived, yicldiJ~g

to ii, mirroring it-and hcn<.:c imitating it bodily (ibid. 45) (see above H1

'Captivatiun', Sect. 5.2). Hut for rhe moment I will approach the sympatheticmagic problem from a different direction. .

Fnlzer's irllellectualism tfcated magic as a form of mistaken causal Thmk­lng. i\nti-Frazeri;\Ils, ever since, have criti<.:ized Frazer for attributing 'causal'

intentions to hehaviours which were symbolic or expressive (Beattie 1966).There is another approach, though, which can be adopted, which is not to con­demn Frazer tor having invokc(] cau~alit~ at all (because magic is, after all,intended to cause things 10 happen) but to rethink the idea of'C'J.use'. Fr:lzer'smistake was to impose a pseudo-scientific notion of physical GlUSe and clTect(cncompassing the entire universe) 011 pf;.lCtiees which dcpend on intentional­ity and purpose, which is precise!) \\ hat is missing from scientific determin­ism. Magic is possible Ix-cause intemiolls rallse eWllts tIJ !lappt'll i1/ lilt' wrillily ofagmls, but this is a diffcrent species of causation from the kind of causationinmh-cd in the rising and selling of the sun, or the falling of Newton's appleetc. For instancc: here before me is this boilcd egg. What has C3LL<;ed this eggto be boiled? Clearly, there are 1\\0 quite different :mswers to this-(i) bc<.:auseit was hC"J.ted in a 5.1UCepan of watcr o\"cr a gas-f1amc, or (ii) becaLL-.c I, off myown bat, chose to bestir myself, take the egg from its box, fill the saucepan,light the gas, and hoilthc egg, because I wanted bfl-aL.-fast. From any pra<.:ticalpoint of vicw, type-(ii) 'causes' of eggs being boiled are infinitely more sali­ent than type-{i) causes. U there were no breaHast-desiring agents like meabout, there would be no hens' egg~ (e\"l:cpt in the South-wt Asian jungle), 00

saucepans, no gas appliances, and the whole egg-boiling phenomenon wouldnever transpire and ne\cr m::ed to be physically explained. So, whatever theverdiet of physics, the real rQusall'xplnflafion for why there arc any boiled eggsis that I, and other breakfasters, i"fOld that boiled eggs should exist.

There is nothing mystical im·olved in tracing the eausation ofe,'ents in one's\'icinit) to intentions or aClS uf willing or wishing pcrformed by oneselfor otheragenG in one's neighbourhood. ThaI is how perfectly ordinary nents do ordin­arily happen-barring 'accidents', and who is ever to know that an accident hasnot been willed by somebod}? Fnzer's mistake was, so to spe:lk, to imaginethaI magicians had some non-standard physical thl'Ory, whereas the truth isthat 'magic' is what you have when you do wi/!lmll a physical theory on thegrounds of its redundancy, relying on the idea, which is perfectly practicable,that thc explanation of allY gi,'cn event (especially if socially salient) is that itis caused intentionally.

The causal arrow between desire and :lccomplishmem reflecls the pmcticalfact thar The more one desires ~omething to happen, the more likely it is tohappen (though it srill may not), Magic registers and publicizes the strengThof desire, increasing the (inductively supported) likelihood thar the much­desired, emphatically expressed, outcome will transpire, as frequently happenswith respect to those ouTcomes we loudly clamour for, 'All events happenbecause they are intcndcd'-'l emph:lrically intend th:1t X shall happen' (Igo'X shall come to pass'. This is nOI 'confuscd' physics, nor is ir deyoid of a basisin social expericm;e, as Malinowski (1935) underslood more clearly than ;my ofhis su<.:<.:cssors, with the possible exception ofTambiah (1985).

102 The Distributed Persoll The Distributed Pasol/ 103

7.3. Volt Sorcery

The symbolic language developed in earlier sections of this essay can be ap­plied to the kind of causality which is involved in Prazer's prime example ofsymp:lthetic (imitative) magic, which is also extensively discussed by Taussig,namely, that form of sorcery in which an image of the victim is maue (often ofwax or some vulnerable material), subjected LO injury or destruction, with theresult that the victim urthe sorcery sutlers the same injuries or is done awaywith entirely. This kind of sorcery is practised in innumerable forms, all overthe world, as anyone who opens Frazer's book is ahle to discover.

This kind of sorcery can be practised in aniconic forms; the ancient Greeks,for instance, used to sorccrize une another by taking small pieces of lead, onwhich they would scratch the name of the 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 ufthe victim inplace of the inscribed lead is interesting, in th:lt it suggests that 'visual repres­entation' and 'binding' (and nanling) are essentially akin to one another. 'Binding'is, indeed, a fundamental melaphor of magico-religious control which we willencounter below. The words 'I bind, I bind' provide a bridge between the lin­guistic side of magic (they are words) and the physical side of magic (they referto a physical operation, carried out on the body of the victim). The action ofmaking a representational image, ofany kind, involves a kind nfbinding, in thatthe image of the prototype is bound to, or fixed and imprisoned within, the index.

Nervousness abuut being represented in an index (a photograph or purtrait)is often discussed as if it were only a foible of innocent tribesmen, who believethat tl1eir souls arc in danger of being stolen away therein. In fact, almosteveryone has reasons for wishing to keep some degree of control over repres­entations of themselves, rather than have them circulate freely. I might beresigned to having my face photographed and circulated, but I do not feel thesame about my naked behind. There is no reason to invoke magical or animisticbeliefs in ord~r to substantiate the idea that persons arc very yulnerable indeedto hostile represent:ltion via images, not just to cruel cariC:llures, but even yiaperfectly neutral portrayals, if these arc treated with contumely or ridicule. Itis not just that the person represented in an image is 'identified' with thatimage via a purely symbolic or conventional linkage; rather, it is because theagency of the person represented is aetual1y impressed on the representation. Iam the cause of the form that my representation takes, I am responsible for it.I cannot disown a photograph of my 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 1 cannot blame himfor the way the picture came out.

The 'magical' aspect of volt sorcery is only an epiphenomenon of our failureto identify sufficiently with sorcerers and their \·ictims, our esrrangemcnt frum

~Indcx-A (volt) ("all"" nJJ"ffi to .. Pr?t~type-l'~ (Vl(llm)

Sorccrer-A lndex-P (volt)~ . .

( ,lnJ\lrC'5

Sorccrcr·A (artist) Index-P (volt) L __--''-'_'-':.:c==--l~ake5..

Prototype-A (\·il"tjm) Som:rer-P (artist)•

appears to

FIG. j·3!1. Volt sorcery and the par(){loxical agen<.:y ofrepreselilation: the victim;s both agcllland pal;cnt

them, not the result of their cnslavement by superstitious heliefs entirelv dif­ferent from our own. \Ve sllfTer, as patients, from forms ofagency t1lediat~d viaimages of ourselves, because, as social persons, we arc present, ~ot just in oursingular bodies, but in e\'erything in our surroundings which bears witnessto our existence, our attributes, and our agency. Volt sorcery is not a moremagical but just a more literal-minded exploitation of the predicament ofrepresemahility in image form. It docs not take leave of the everyday world,in appealing to some occult foree, some magical principle of causation; on thecontrary, it unites cause and effect all too closely, so that the causal nexuslinking the image to the person represented is made reversible-the image canexercise a causa] efTect, in the opposite direction, over the person. The modus

operalltli of volt sorcery can readily be expressed using the diagrammatic con­ventions introduced earlier, as in Fig 7.3!1, or:

[[ [Prototype-A] ----,) Artist-A1----,) Index-AJ~ Prototype! Recipient-P.

The verisimilitude, so til SIJeak of volt sorcerv 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 recipienr, whose injuries stem from the injuriesthat the index has received. The victim is ultimatelv the victim of his ownagency, by a circuitous causal p:lthway. Vulnerability ~tems from the bare pos­sibility of represent:ltion, which cannot be avoided. Sorcery beliefs endure, andarc highly explanatory, because vulnerability to sorcery is the unintended con­sequence of the diffusion of the person into the milieu, via a thousand causalinfluences and pathways, not all of which can be monitored and controlled.Frazer himself noted that image-based sorcery is c10sel \' allied to the mher kindof sorcery which depends on exuviae; hair, ~ail-dipp;ngs, food leftovers, ex­creta, and tbe like. Often, []le volt is rendered more effective by the incorpora­tion into it of the victim's exuviae, so that imitative magic, based on (visual)similarity, is allied ,,·ith the other main kind of magic, based on contact. Onceagain, to describe this as 'magic' and LO imagine that such sorcery is based on

IIH 1h~ Disfrilmled Person The rJislnhu!<·d Pawm 105

occulT principles (or symbolic rituals) is quite misleading. There is nothingtranscendental about the kind of causality involved in exuvi,le sorcery, thoughoften sorcerers may enlist spiritual helpers in pursuing the objects of theirhatred. Exuviae sorcery works (or seems likely to work) because of the intim­ate causal nexus between exuviae and the person responsible for them. TI1Cseexuviae do not stand metonvmic,llly for the victim; they arc physic'llly detachedfi-agmenls of the victim's 'distributed personhood'-that is, personhood dis­tributed 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 of the prototype, and the index as a(detached) pllrt of the prototype. We arc not accustomed to think of images(sllch as portraits, etc.) as parts of persons, limbs, as it were. In terms of thesemiotic thcory of representation, nothing would be more erroneous than toimagine that the substancc of a sign (the visible or audible sign 'dog') were partof any dog, or dogs in general. Dut with indexes it is not the same as withproper signs. Ahduction from an index does characteristically involve positinga substantive part-\vhole (or part part) relation. Smoke is a kind of 'part' offire, [or instance. i\ person's smile (the cheshire cat excepted) is a part of theii-iendlv person it betokens. r;'rom this point of view, it is not senseless tosuppo;e that Constable's picture of Salisbury cathedral is a part of Salisburycathedral. It is, what we would call, a 'spin-off' of Salisbury cathedral.

74- The t"pitllrea1l Theol)' of 'Frying Simulacra' as Parts o/the Body

The convergence between images of things and parts of things can beapproached from a philosophical angle. Yrji) I"lirn (I900, cited by Frazer I980

and discussed also by Taussig 1993: jI) 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'. Him writes:

For it is evident tlwt a philosophical doctrim:, if it fits in with the f,lctS of primitivesuperstition, may be explanatory of those vague and latent notions II'hi,c~, withom lo.gicaljustification or systemati(;al arrangement, lie in the mimi of the maga:l.an and the. Idol­ater. Such a doctrine is prescllkd to us in the familiar emanation-theones, aceordmg towhich cvery image of a thing constitutes a concrete part of rhat thing itself. "~c~ordingto the ckar ,1llJ svstematic statcment of this doctrine given by the old EpIcureanphilosophers [I mili provide the rcln;all! quo/a/iilufro/1l Ll~[rrli/ls.;'elolP] sha(~ows, rdlec­tions in a mirror, visions, and even mental representatIOns of dtstmtl ob)e(;\s, arc ,Illcaused bv tbin memhranes, which conrinually detach themselves from the s\lrf.~ces of,Ill h()die~ and move onwnrds in all directions through space. If there are such things asnecesS,lfV misconceptions, this is certainly one. Such general f.tcts of sensuous expen­en(;e as ;·dleelion, shadow, and mirage will natunllly appear as the result of a purely

material decortication-as in 'I transfer piuure. I foil' near at hand this theon' may liee,en LO the modern mind appears from The curious fau that such a 111an as riall.a~ fellhack on it when attempling to explain the newly invented Daguerrotype, that most mar­\cllous of image-phenomena. (I lirn 1l)00: zo;lj-4; cf. on phOTOgraphs as indexes of thercal presence of persons, Barthes 1981 cited and discussed in Freedberg 1989: ch. 15)

The doctrine of emanations comes directly from Epicurus, but the mostfamous, and [or my purposes the most interesting, statement of the doctrine isprovided by Lucretius, who writes:

! will attempt to lay before yOll a truth which most concerns ... the existence of mostthings we call the idols [SII11I1/I1I1'I/; in Greek, ~id(llill of things: these, like films peeledoJT Ii-om the surface of things, tly to and fro through the air. , . I say then that picturesof things and thin shapes arc emitted from things off their surfacc, to which an imageserves ,1S a kind of film, or nallle it if you like a rind, because stich image hears anappeanmcc and form like to the thing whatever it is from whose body it is shed andwanders forth.This you 111ay learn howevCT dull of apprehension froT~ whaT follows.[.\lany visihle obicetsl, ... emit bodies some in a state of loose difTllsion, like smokewhich logs of oak, heat and fires emit; some of a closer and denser texture, like the gos­S,1111er coats which at times eicades doff in summer, and the films which (;alves at theirbirth cast from the surface ofthcir body, as well as the vesturc which the slippery ser­pent puts olramong the thorns; for often we see the brambles enriched with their Oyingspoils: since these cases occur, a thin image likewisc 111ust be emitted from things offtheir surface. (De RerulIl NII/ura + 26 If. tmns. ,\luIlTo, pp. -1-4-5)

Lucretius attributes the flying simulacra of things to a kind of internal jostl­ing within objects, which causes the minute bodies 'in the front rank' to bedischarged from the surface and to fly outwards. The simulacra arc physicalthings though, and we sec objects because simuhlcra enter our eyes and we canfeel them 'since :1 p:lrticular fibrure felt by the hands in thc dark is known to bethe S:lme which is secn by the bright, light of day, touch and sight must beexcited by quite similar causes'. Lucretius discusses a number of optical phe­nomena, 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 physie:ll forms of diffusion into the amhience, particularlysmell and smoke, as well as the shedding of skins, rinds, and films from thesurt:1Ces of things. I shall haYe occasion, later, to return to the conceptuallinkagc between smoke, smell, skins, and visual appearances, which KUchlerhas identified as key elements in the ideology :lssociated with a particularlywell-known 1Vlclanesian art form, the Malangan carvings from northern NewBritain (see below, Sect. 9.2). For the moment, though, I am interested in I lim'spoint th,1t if 'appearances' of things arc material parts of things, then the kindof leverage which one obtains over a person or thing by having access to theirimage is comparable, or really id~ntieal, to the leverage \\hich can be obtainedb) having access to some physical part of them; especially if we introduce lhe

Tht {);Slr;bllud P,'rstm nu D;slrlblllt:d Pl'rJlJII lOi

nOlion that persons may be 'distributed', i.e. all their 'pans' are not physic-.lII}auached, but arc distributed around the ambience, like the discarded 'gossamerl.uat.'i of cil.'3das' in I.ucretius' memorable instance, which arc both images andparts of the living crC.1ture.

7S From SorcerJI!o the Cult ofImages

1 (Urn noll' to an ethnographic cxample which provides a bridge between VQIrsorcery, surely a discredirablc practice, and the wOTflhip of imagL";; in the religiouscontext. Volt sorcery provides:l model for understanding the worship of image-;in general-indeed for 'objectification' in religious contexts generally. Thematerial and the idea~comes from Alain Banadzan's rcmarkable reanalysisof the religiolls practiu.."i of ancient 'rahiti, and the deploymem of images therein.

It may ha\'e been noticed that the \riClim of volt sorccry is involved in anim-oluntary process of rxduwge. This arises natur;l!ly from the fact that heappears twicc in thc formula, once as the comributor of somcthing (his appear­ance) and once as the recipiem of something (injuries matching those sufferedby the \'olt, the index). He is an 'involuntary' agent; voluntary agem.-y lies withthe smcerer-who may, of course, have been justifiably provoked.

The origin.llit}' of Babad:-..an'" account of Polynesian sorcery and idolatryarises from the subtle \\-ay in which he has obsencd both of these werc varia­tions on the well-known, bUL oftcn tantalizing, explanation of the process ofexchange provided by Ranapirl, The Maori intellectual, to Elsdon Dest (cl'.Mauss 1954; Sahlins J()7+).

I will now spc;lk of the 111I11, amllhe ceremony of rrhangll; hau. That III/II is nol lhe 1/111I(wind) that hIO\ls-nOf at all. I will C'.lrefully explain to you. Suppose that you POSSl.'SS

a cITtain article, and that you gi,e th;lt :lrticle to me, I,ithout price. We make no bar­brain O\'er it. Now, I gi\ e that article 10 a third person, II ho, after some time has elapsed,decides to make some retorn for it, and so he makes me a present ofsome article. :'\ow,thai artide he,: gi\'cs me is the IrUIi of the article I tirst recci\t~d from you and then gaveto him. The goods that I rccci,'cd for that iu_-m I mLL~t hand mer to ~-ou. It would notbe right for me to keep sueh goods for m}'sclf, "hethe,:r lhe~ be desirable items orotherwise. I must hand them o\er to ~ou, because tho:y are a IrUII of the artiele you b":t\'eme ..

I I\-ill explain something to ~ou 300m the fiJrt:st 1/(11/. The mallr; was plal't:J orimplanted in the foresr hy the II/lillI/gil. It is the mllllri that causes birds to be ahundantin the fotest, lh;lt thcy be sh1in and taken by man. These birds arc the property of,belong to, the 1IIII/II'i ... l!cnl;e it is Said that offerings should be made t~ the ':1111 ~f thcforest. The IIJlllwgll (priests, adepts) cat the offering: because themaIlYllsthe1r5:n.WaSthev who loe-.Ileu it in the forest, who C;lUsed it to be. That is why some of rhe birds(:(}(;ked ;\1 Ihe saered tire :Ire SCI apart to be eaten b~' the priests onl~·, in order lh;ltthe 111I11 of the forest-products, and the ml/llri, ma~ return ag:lin 10 the forest-that is,to the ml//lri. (Best 11)0<): .f39)

Babadzan explains that the exchange process described in Ranapiri's textinmh-es thrce participants, the priests (lolllll1ga), the '/,au (If the forest', and thehunters. The 'hau of the forest' can be glossed as the 'principle of increase' inthe forest; its fertility in other words. The priests make offerings of mallri tothe hnu of the forest. The hau of the forest responds b) providing the hunt'crswith birds to capmre. A porrion of these birds must be returned to the pricsts.The offering made by rhe priests and referred to as maur; arc 'fertility stones'.The mauri forge the link between Ranapiri's famous text and our theme in thischapter, for mauri arc aI/iconic ;t/ok They are index(.'S, in other words: they areobjectified repositories of the spiriT (of increase) of the forest. They mighT justbe special stones, but Rest also tells us that sometimes they took the form of'a hollow stonc, in which hollow would be placet! 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 mOllri was createdby immuring a (living) li:-..ard within a hollow tree beside a bird-snaring site(ibid. .J37). The significance of thc hollowness of mallr; idols will emerge in duecourse (below, Sect. 7.ll).

Let us return to Babad7..an's exposition of the ritual sequence, in which, inexchange for the birds they captured in the fore."t, Maori hunters were obligedto recompense the priests (tohunga) with a portion of the game they secured.

The magic stunes and the birds the hunters capture arc thus, rilUally, nne and thc samething, one ami the same II/Ullgil (gift) that the prieSl5 'gi\'e' to the hunters, via the inter­mediary agency of the forest. (Babadzan 199.1: 6.f. my tf;I!lS.)

In oLher words, the forest is obliged LO give birds to the hunters, nCC:luse its life,its capacity to manifest productivity and fertility, is not its own; it has beenplaced there by the priests. 'The flail of the forest, which meanwhile is, inMaori thcory, the very principle of the productivity of the forest, is I/'IIS COI1­sideull as ti,e pass;VI! agent iff a tranrmissiun of wllich lire pril'sts are Ihl' primemOI~rs' (ibid., italics in original, my translation). But there is a paradox here.How GlO a forcsl, 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 "ice \ersa. What is meant is that the forest is passi\'e in relation to

the priests, the prime movers, bUl not that it has no intrinsic agency at aiLThere can only be a mauri of the forcst, a physic'll objectification of lhe pto­ductivity of the forest, because the forc.'it is (polenti;llly) productive in itself; itis this 'agency' which has been co-opted by the agency of the priests.

The fundamental similarity between rhe situation just described and \'oltsorcery will perhaps be beginning to show itself: the mallri (objectified fertil­ity) created by the priests and buried in the forest is both a represent;ltion ofthe prodllctivit} of the forL'St (t'x aI/it!) and a cause of the forest being fertile (expost). Because ex a1lte fertility C;1Il be represented-Leo objectified in an index­i( comes under the comrol of those who control the index, the priests. The

priesrs, in other words, make an index of the productivit), of the forest and thatmakes the fOROSt productive. They arc rewarded by the hunters, who recci,·cfrom the forest the fruit of irs producti\'it~,. and who return it to the priests.The magic, so to speak, jlL<;t has to .....ork. because the index of producti,oity iscaused by the productivity it causes; there is a perf(''Cl, but invisible, circularity.The objectincation of the productivity of the forest diners from volt sorcery,of cour~c, in one essential respect, namely, that the mauri represents the forestas prosperous and flourishing, whereas the \'olt represents the sorcery victimas injured, or even dead. But the mt.'{;hanics arc the same; the pric."rs makethe forest flourish by representing it as flourishing; on Ihe other hand, thevery same objecrification could be used to injure and kill the forcst, were itsubjected to abuse. In fact, if enemies managed to find the moun of the forest,they would destroy its effectiveness by reciting productivity-negating chantsover it, bringing death to Ihe owners, for now these &1.me mOllr; would object­ify miser)', and cause it. This is the (:qoivalent of sticking pins into a volt toinjure the victim represented as injured.

What is most interesting of all to note, though, in this conne!,;tion (DabadzanHJ9T 61) is that III/II (fertility-principle) objectified in the mlluri, is also theword used to refer to the exuviae used by a sorcerer to cnsorcell his victim(citing Tn.:gear 1891: 52). We now sec wh~' hollow I1Il1uri had 'locks of hair'placed inside them. Babadzan suggests a most satisfactory e.xplanarion of thcsynonym), between sorcery exuviae and the principle of fertility: both involvegrowlh. Exuyiae arc parts ofthe body which ha\·c grown and become separate;this particularly applies to hair and nail-clippings; indeed, where adults arcconcerned, these arc the most obvious manifestations ofgrowth that the humanhody provides; and even if hair remains uncut, it still f.'ills out and separatesitself. Exuviae represent 'growth' because they :\I'c, so to speak, continually'h~rvested' from the living body.Just:ls \\"C harvest ollr hau, whenever we havea haircut, so when the hunters entcr the forest 10 'hal'vesl' the bird~ there, theyare harvesting the exuviae (hau) of Ihe forest.

Exu\'iac sorcery is possible because of the fact that as the bod}' grows, itsheds its parts, and these become distributed around the ambience. Hl.Te onemay rCClll Ihe remarkable resonan(:c between Lucrelius' instancing the skinsshed by snakes and cicadas as prototypical simulacra or 'idols'; for these are,prl.'t:isely, exuviae prodoced by growth and distributed around. In f.,ct, theEpicurean theory so.w lhe generation of simulacra as a 'growth' process-the'shedding' of ephemeral skins from all things induced by a kind of 'pushing'from within.

The IIl1l1tri, in irs guise as a fertility stone rather than as an itcm of humanexuviae, objectifies the growth of the forest because il is, conceptuall~', some­lhing produced by lhat growth; il is lhe exuviae of the forest, which fallsinlo the hands of Ihe priests, who use it in the 'white' magic of prosperiry­induction rather Ihan the black magic of sorccr~ and dearth-induction.

Physically speaking, mauri fertility Slone." could take a variety of forms as wehave seen. Forest mauri were aniconic, but those used 10 pro~O{e the ~owthof swcel potatoes were can·cd as anthroponnorphic images with open mouthssometimes with their hands in a position suggesling [he placing of food in th;mouth. Ensuring that the poeple ate, they were shO\\n cating-and they wereof course fed themselves, with offerings of first-fruits and so on. Fig. ;.5/1 is a\'l."(Sion of the figure Babadzan provides (0 summarize his general argument.

7.6. Detorticalioll {{ut! fhe E:ryhfl1lge uj1nrlexes: Tahitian To'o

The Tahitian equivalent of the Maori JIIallr; slone took a varicl\' of forms also.The productivity of fisheries was controlled by a rype of Ji'i -(fiki = 'image')called puna. These 'fish' smnes were elongated, like fish, and certain examplesha\·e fish-like gill slit.. on them. There arc dCS(,TIptions of the rituals associatedwith puna which make it dear that their use was entirely analogous to 11Ifluri.

For agriculrural purposes, Babadzan cOl1\'incingly argues, a different t) pc ofstone was used, carved with anthropomorphic features,like the anthropomor­phic /Iltwr; buried in S\\CCt-potato fields in :'\few Zealand (ibid. 75-82). Thesegrowth stones were actually believed to grow, though only very slowly. Notmuch documentation rhat is relevant to agricultural rites in Tahiti has unfor­tunately sun'ived, but it is safe to say thai increase-ritU'lls comparable 10 themuch beller documented Maori ones were practised there (ethnohislOrical1y,Maori civilization and language have antecedents in the Society Isl:mds, of\\hich Tahiti is the most important). Perhaps the most important poinll'O bearin mind is that the Tahitian word lIIpu, which hasically means 'growth' alsomeans, like 111111, exuviac, usable for sorcery purposes. Tupu is the Tahitianeqoi\·alent of hall, in other words.

The il1(erl.·~a of the Tahitian material discussed hv Bahadzan lies in a differ­ent direction than the merc replication of the l'vlaori ~attern ofagricultural rilesand the concepts of hllll, II/flllri, etl.:. M:lOri societv was effectively decentral­ized throughout, but in ancienr Tahiti there was a l;roro-St:lte with; centrJ.li7.cdcult, the cult ofOro. Consequently, in relation to the Tahilian material, we areable, as it were, to pass from volt sorcer), (;l pri\·ate alTair) to public "orship ofthe most majestic and awe-inspiring \'ariel)', involving the state god, the 'godof po\\ er', Oro.

108 The Distrihuted Person The DislrihUied PrfJon

Pr1ests rnaufl L_""'_"'_"_-, _birdS LJ~~'(create maoo) -----+ (Pf~oIltle Forest

t hau 01 blrds t hal) 01 rnal)'; I

109

FIG. H/I. Thel:irculuil) ufproductin, t.'tch~ngesbetwccn index(mauri) andprowl)"pe (!tau) arlerBabadzan t99.1

1I() The Di;;tn'llIIled I~[,/,sol/ 111

roll;_ 7_6}1_ Ti'j in iL~ bcnign form asa '-ole,mic gt"uwlh Slonc_ Sallrer: aft·crI'ierre '-olca"i\lue)o em., SociClyIslands. i\tusCe tie I'Homme, Pari~

Unlike li'i images (see Fig, 7,611 )----such as fish stones and anthropomorphiccarvings, used in agricuhure and sorcery-the objects at the ccntre of the Orocull, and the cults of lesser gods, anc~1:ors of important kin-groups, were nOIanthropomorphic, though they were (in a sense) representational, and theywent by a different name, 10 '0. The word 10'0 means a 'prop' or s1'J.ff or pillar.Thc to '0 represent, mythologically, the pillar placed by the creator god to holdup the sky and pr~r\'e the ao (the world oflight and of human life) from theencompassing powers of night, darkness, and di,'inity (the po). It is a mootpoint whcther the to '0 deserve to be considered representational (i(:onic) ornon-representational (anil.:onic). A.s elongated billets of wood, slightly thickerat one end than at [he other, they might be regarded as 'realistic representJ.­lions' of pillars or props, which L'i \\ hat they are called. One could call them'iconic' indexes of pillars, in that they refer to the mythological pillars whichhold up the sky without actually being these pillars. On the other hand, the)represent, aniconicl1ly, gods who havc anthropomorphic attributcs, Oro wasnot a pillar, as such. Rather, as the pillar which keeps heaven amI earth ;\paTl­and thus connects them-the fIJ'u is invested with the god's presem;c hy I irtueof contiguity rather thall resemblance. Dabadzan himself suggests that the 10 '0

is lonnless because lhe god presides Mer the origination of everything thaIpossesses form; as the givcr of forms to alllhillgs, the god is himself formlcss.This deduction is well supported in the Tahitian ereatioll chants (Ilenry 1928;cf. Gell 1993: 124 fT,). [n this case, the {o'u is agaill all iconic image, a formlessim:lge or formlessness, or all thillgs ill Ita/II !wI(Cllt!i. Here, it seems 10 me thatthe supposed cotl(nst between iconic and alliconic representation comprehens­ively breaks down. Thc lu '0 arc wholly iconic and wholly aniconic at the sametime. What is quite il1Jlrguablc though is that they have a prototype, and thattheir prototype is the person of the god. They arc also, like the humblest volt,open to manipulation by human heings. Through them, 01'0 himself, themightiesl being in the Tahitian unil'crse, could be rendered a 'passive agem'

via represenl':ltion, just like the forest hau in thc previous example, or the pas­sive victim of volt sorecry.

Pursuillg this theme, we arc able to diSCO\'cr that the primary ritual pro­cedure in which the to'o were imolved, \\":IS a ritual of decorticmioll, of thecontrolled production of dil ine eXll\'iae, Following Uabadzan's analysis of theexisting sources, I will now describe this ritual sequence. .

The primary 10 '0 were embodiments of Oro. The Tahitian political systemrnoked around asserting comrol over the /lIarae (temples) of Oro; whoevercontrolled the 10 '0, controlled the country, and ri\"al chiefs fought binerly overthem. The important 10'0 stood at the pinnacle, so to speak, of a polit~ ofimages, and social rank among human beings was strictly and precisely corre­lated with the 'rank' of the 10 '0 (and lower down the scale, li'i, 'sorcery idols')in their charge, Social rank and polilical »Oller I~ere gi"en regular and fonnalexpression at 3 ritual called po 'inJua, which, translated, mcans thc 'wrapping ofthe gods' (pa 'j::: wrap up, tic, bundle). (On wrapping in general see Gclll993.)

Only during Ihis ritual were 10'0 revealed 10 view, and even then, only 10

the most important and ritually protected priests and chiefs; the sight of a10'0 would insnmtly result in the death of a lesser person, so filled were thc)with mana (a quality called ra '0 in Tahitian). To protcct thcm from view, to '0

were at all othcr times tight!} wl'Jpped in bindings of sennit cordage and tapabark-doth, sometimes with roughl~ delineated facial fC'.ltures and limbs W(l\'cnonto the outside (Fig. 7.6/2). Thc more import.1nt ones were kept in specialc(llll:ainers for stonge and tl'J.nsporl. The ritual of pa 'ialaa consisted of theassembly of all the images in a district at the lIIorae, for the purpose of therenewal of their oUler wrappings, espccialJ) the outer wrapping of the primary10 '0 of Oro.

I m..'Cd not describe the ceremony in any detail. Essentially, it consisted of aprocession of the gods into the /lIf/rae, led by the main god (completely hiddenin its ark), wilh the lesser onts follo\\ ing in order of precedence, followed (andprotected) by 'sorcerers' bearing nOI f/) 'u, but li'i, sorcery images (Fig. 7·of3)·LI.,'sscr gods and li'i 'sorcery' images, Ellis tells us (and there is a sole surviv­ing example in the Dritish J\Jluseum to back lhis up) were not in the anieonic'pillar' style, but were hollow anthropomorphic images, with an intcrnal cavityfor holding sacred feathers alld/or sorcery exuviae (Ellis ,831: i. 339). AUwould assemble in and around the II/firtlt!, according to their stalinn. In.l separ­ate sanctuary, well away from public I,jew, the tv'o would be divested of theirwrappings, oiled, and laid OLlI ill the sun. There thell ellsued an importantexchange sequence.

The most symbolically significant 'currcncy' of pulitical authority in Tahitiwok the form of the feathers, the moSI impoTl'lllt beillg red ones. \\Tearing rcdfeathers was a sumptuary privilege of the highest chiefs, the lllfirll lira (redfeather girdle) was synonymous with occupancy of the paramount chicfship.

112 nit, ViSlribult:J Pt:r.<i!ll nit' Dis/rilll/Nt! Pt:rslm 113

, In s,mo:l Ih~ ...·.1 fe,..hers ,.-o\"en inw the n...",t prestigious fine m~l~ furmillil: chien~· dowrie,.; "e""asso...i:ll..... ,,;Ih lhe h~menc:l.l hlood of the 'lry:in "'U1"". 1hc l'rcdKlineJ reproducer of ch;cn~ (lOwer.

F'e;uhers, IlI.'Ctllcss to say, arc the exuviae of birds, .Illd birds hnc, C\'cfp\hcrcin Polynesia (.IS so often elsewhere) hC:l\'cnly associations. Red fcathers afC

probably :llso associated with blood, another exuvial substance anti index oflife, growth, :llld reproductivity,' Oro was particuhtrly :ISSIl(iatcd with bOlllfcathers and redness. The prim.try :lct of homage to 01'0 was rhe presentationto him of red fC:lthcrs, which would be woven into his senn;r wr:lpping,auachcd 1'0 it by cords, or simply wmpped inside the co,·cring.

FIG_ I _6/:!_ Too wilh Orothc goo hidden from ,'i"..~nd wrapped in binding.,;.Sourer. Museum orArchacolog~ :mt.!AnthmpollllY, Cambrid~.Z606J

FIG_7.6/]. Sort-ery Ii"; with hollol\' hJcl for fcathcn;/c.\u\"iac.Sourer. The Brilish ;\1u.'iCum.I..ondlln. 1\1:\1024}679

When all were SC3lcd, the high priesl opened the ark and tool.: out the dreaded image,and as he um:o\"ered it upon the mal, Ihe others all unco\ erctl theirs in unison ... Theminor b'lKis thcn e}.~J, with Iheir wrappers folded umler them, remained in thehand~ of thcir o\\ners, facing the in:a jl T("J.d, for prescntlltion ro the tulelar god (i.c.Orol ... When Lhe image of Ihe lutc!:lr god was reH::ak-d from the profusion of redand ydlo\\ fe~tlhers Iring upon irs man) CtJ\'erin~~ ... Then follo\\ed the prcscnt:uionof the minot gods by Lheir owncrs in lheir proper lurns, with offerings of new Urtl

(red feather] amuleL~ and loose fealhers, which wen:: gi\t~n through the high priestLo the mtclar god in exchange for somc in his possL"Ssion ri.c. which had just hl"Cnunwrapped from thc scnnit wr3pper of OroJ. This act was C'Jlkd lari/tlt/ra'lI-allltl

(the god's !':xchange) and was supposed Lil add new power from the grealer god lU thelesser ones. The Iishermen's gods were presented IasL hec;1Use they wen: from the sell,(Henry I92R: 166-j)

Babadzan pro,'ides the following schema ohhe feather exchange: new, non­sacred fl.."arhers passed from the lesser priests to the pricsls of the primary 10 '0

(the 'main' Oro) and the feathers which had prc"iously been in contact withthe primary 10'0 were passed back to Lhe inferior 10 '0. In this way a portion ofthe sanctity of the primary 10'0 WaS distributed [0 the assembly of lesser to '0 inexchange for a rribUle paid in the form of new fl.."athers. a 'natural' prOOucl, asil werc, potentially embodying po\\er and fcrtility, btl( not }et able to generateit. Only fealhers which, as Dabadzan notl.."5, had been intimatcly in contaCI withthe primary image of Oro for some lime, had Ihis gencrati\'e property (1993:Il6). The feathers had to live and die with the god, quite literally, beC,ltlSe ilseems that to become divine exuviae, the primary god has to 'die' alld scalierexuvine back into the world. This, Rabadznll thinks, is the significance of Lhesequence as a whole, that is to say the unwrapping of the gOOs followed bywrapping them up again. enwrapping brings the god itHO the world, wrappinghim up alr.tin sends the god back to lhe nether world, the po. Bm beC;1usc thegod has died but left some!.hing behind (his fe-.tlhcfS. his exuvi:le II/PII) thepower Ilfthe god is also lefl behind-in the hands of the chief" andlhe priestsofOro. This inference is based on the fact that !.he Tahitian mortuary ritual forhigh chiefs imoh'ed the drying of the corpse in the sun, on a platform, followedby wrapping (binding) and decor-uion with rcd feathers. This 'HS to keep Ihemighty dead under control, safe!) dead, and incapable, because of the massi\'ebindings of hark-doth emeloping them, from harming the li\'ing. Such tre:ll­mcnt (wrapping and decoration lI'ilh fcathers) was reserved for corpses of pow­erful chiefs ,md for the to '0, so Ihe implication is th,H the 10 '0, at the cllnclusionof the pa 'iaillo ceremony, is being treated :IS a 'powerful corpse'. 'I'he bindingup of the /1)'11 places Oro in the 'patient' position.'

• Reldled 10 Thi~ idt:"~ of·bin<.lio;:· Ihe gOO;~ !he procT;"". n hich n'~< ("mmon in ~rlli<.juiT~. ofre~tro;n-

ing The iOl~gC'l of lhe gOOs in lempk." nilh ch~i and Ol~fl.lck.,,;, 10 pre'cOl ,hen' frum cs<."";lpiny: andlr.lIl'lferring Ih.,;r l,rolcc,;on 10 ri,~1 <:ill.,.. (.·f. F -.:Jocrg I<)&,; 7"~5l.

114 The f)istribllud P<'n,m Tht Distribuud P"mm 115

The cult of the gods In the form of 10'0 and the accompanying fe-.uhcrexchanges can be seen as a kind of ,'astly magnifit-d. exercise in \'olt sorcery.Ostensibly, Oro rules the uninrsc, but in practice his intervention in humanaffairs is controlled by the chiefs and priests, via indexes which arc his parts,his cxu\'iae, and the binding oCthe primary index of his person, Ihe 10'0. It \\ illbe recalled fhat the ancient Greek sorcerers wrote on the little Ie-Ad plat~ rep­resenting their victims, the words 'I bind, I hind'. In the cultoflhc 10'0 we findthis 'binding' of the god literally enacted, but the implications arc the same.The prototype of an index is bound to the index by resemblance, and is thussubjected to control. "I'he to '0 \\ efC lhcmscl\-es tightly bound, and their powerdi£fused into these bindings, eSIX'Cially the feathers, which then bct-ame thecurrenc)' of political control,

~ This idea can perhaps be givcn a more genernl statement, so as to apply to,i )all religious art. The great monuments that we hne erected to God, the great

" basilicas and cathedrals, are indcxt'S from which wc abduct God's agcnq over':$( the world, and over his mortal subjects, who have striven and laboured to

, please him, and have left these milssin; shells (or skins) in tl1eir wake, withinwhich the faithful gather LO worship the ultimate amhor of all this magnific­encc, Such, at any rale, is the orthodox view of religious magnificence, whichrepresents basilicas and c-J.thedrals as 'offerings' to God, who is all-powerful.However. it must also be rL'eognized, first ofall, that God is n~ rt':llk powerfulat all unless his power is apparem in this-worldl\' indexes (behayioural ~nes,

7 orL III the present l;:lSe:tilaterial.2ncs), The basic ~Iiltonic paradox thar Go~nsat r1Sk Irom hIS own creation, simply by hning distributed himsclfin manifoldforms (including Saran), applies lO such material indcxes ofGod's greatness inrhe following way, Humanity has a lien on God because his objL'Ctification is intheir hands. Even if God is the ultimate author of his resemblance in the formof magnilicent structures and works of art, it l'emains the case th.H, at a crit­ical point in the sequence of causes, instruments, and results, human agencyis eSSClHial. Sincc, in this world, God's presence is inherent in these works ofhuman agency, he is bound to hum:m purposes, the this-worldly prosperityand other-worldly salvation of his ostensible servants, rather than to purposesentirely his own, His agcney is cnmt'Shed in ours, by virtue of our C:lpacity to

rna);c (and be) his simulacrum, With respect to our god, we are in just the sameposition as the Maori (ohtmga with respect (() the hall of the forest.

Of course, the fact that we have tr:.lpped God inside his likenesses does notmakc all religious 'Ictivity sorcery, as such. The homage paid (() God in thebasilica is not destructive or malignant, but it docs m11 ke him rhe 'passive agent'of essentially human dcsigns, jmt as the homage paid to the to '0 made Oro thepassi\'e agent of the chiefs and high priests. J ha\e no doubt that Christian the­ologians would ha"c no difficulty in refuting such all imputation in a manncrconvincing, al least. 10 themseln;s, but rhe logic of the situation seems to me

inescapable, The papal title 'Pontife.'( MaximlL'>' (J\'L1V5 II PONTWEX MAX[\IV5

blazoned ovcr the apostle of 5t Peter's in Rome) attributes to the Pope thepower to build bridges betwL"Cn earth and heaven, The analogy between thepapal 'bridge' between earth and heaven, and the Tahitian to 'o-the staff Otprop whieh lceps heaven and earth apart (but also in t"Ommunication)-seemstOO remarkable to pass without comment, And in St Peter's one can certainlydetect the clear implication that God is the exchange-partner of his mor~important subjects such as Julius II if not the mass of his worshippers of lowestate. St Peter's is the bridge; but the point is that the making of the bridgehas been attributed unambiguously to the Pope, Julius II, his predecessorsand sut'Cessors. He, in exchange terms, is the primary donor, the holder of thekitoum, the 'unencumbered valuable' (Munn; sec below, Sect, 9.3) which issent out to find its match, thc valuable which can be measured against it andreturned for it. St Petcr's, as a gift-objt'Ct and an index of human agency, eli­cits a responsive counter-gift, which, parddoxically, is Sl Pcter's itself, im'estedwith divine power now a\'ailable to m::mkind, like the feathers which, given roOro, return from Oro as his embodiments.

Howevet, let us not dclay too long over such .m.dogies, which may beregarded as unconvincing, or even offensive. I now want to turn to the wor­ship of anrhropomorphic images, among which neither the /0 '0 nor St Peter'smay be counted, at 1~.lSt, not at first glancc.3 The literature of idol-worshipis, in the main, profoundl~' unsympathetic to this practice; it is almost as iflearning to read and write disqualifies one from engaging in this practice withany enthusiasm. This rather supports the notion thar there is a 'great di"ide'between the essenrially non-sensuous mode of literare thinking and the sensu­ous, participafOry mode of pre-literate thinking, Howe\'er, there h.ls been areaction against the theory of the 'great divide' in recent years (Parry 1985) andthcrc cert,linly exist numerous literate image-worshippers in the world today,and ancient, literatc, civilizations whose religious pracrices centre on the pay­ing of homage to images, sueh as Hinduism. The idea that only the uneducatedor 'primitive' worship idols of stone, wood, and mctJ.t fashioned to rt'SCmblethe human form, is a consequence of the com-ergcnce between anti-imagisticforols of religiosity (such as Judaism, Islam, and certain forms of Christian sec­tarianism and Protcstamism) and the rise of a more generalized religious scep­ticism, which has ancient antecedents, Indeed, wherever religion exists, it isprobable that scepticism also exists, whether or nor this is expressed in public.Certainly, anthropologisTs have encountered innumerable examples ofsccpti­CiS111 concerning the etneacy of rites among the illitcl'ate and uneducated, sothis eannOI be auributed to literacy or the rise of 'science' atone.

, But cf. \\ ittLower on anthropomorphic c!cmCl1l'; in Rcnai""'"\tt thrones of archi'CClural propor­tion, AcL~'1'n\.1.non Ikmjni's p;:v~-". ClC_

116 Tht l>islrih/lII!J Pemm TI,,· DiSlrihlllal Ptrsf)7/ 117

7.7. Darshan: Wil1lessillg as Agency~estions of the cultural foundations ofbdicf and scepticism nCLod not detainus but it is still enonnouslv difficult for Westerners and non-believers to

e~pathize \\jlh idol-worshi;pers because of the bombardment with anli­idolatrous propaganda which we ha\"c experienced from the ,'cry moment \\cbecame conscious of such things. Perhaps the moSI insight-provoking aa.:ountsof image-worship belong to the literature of Hinduism, because followers ofthis religion arc among the least self-conscious in showing devotion to images.A convenient introduction to the copious literature on image-worship in Indiahas been provided by Diana Eck (19RS). Worshipping images is obtainingdarshan from the god, a particular type of blessing (onveyed through the eyes.D'mlul1I is something given by the god, a mode of the god's agency in the world,and the won.:hippcr is a patient (the Recipient-P, in terms of our scheme).Living human beings c.m give Janka", as well as gods; a guru gives Janka"when he or she makes an appl.:arance before a gathering of disciples (sec Babb1987), and the same is true of an important politician appearing before anassembly of supporteno, who have come to see, as much as to hear, their leader.Dorsltan is a gift or an offering, made by the superior to the inferior, and itconsists of the 'girt of appearance' imagined as a material transfer of someblessing.

It seems to me that Jorshall is essentially similar to the other mode in whichdivine blessing/personhood is distributed in India: the distribution of sacredfood, prasad, which is consumed by the god's devotees (of course, there areChristian JXlrallels to this as well). Prasad is often conceptualized as the 'foodleavings' (jl/flla) of the god, in a manner absolutely analogous to the exuviaeused in sorcery. According to Eck, the conceplualization of darshan is closelyallied to the role allotted in the Hindu tradition to the eye as an organ of inter­personal transactions.

Darshall, considered as a mode (If divine agency is thus intimately connectcdto the concept of the nil eye. Di\'ine idols, religious gurus, and politicians ofrenown transfer blessings via the steady and penetrating gaze with which theyirradiate the assembly o\"cr which lhey preside. This is, so to speak, the posil­i\'e 'while' aspect ofevil~yesorcery, which more imperfcct beings transfer viatheir me:m, envious, and ill-intentioned looks. To place oneself before the idolof the god, therefore, is to lay oneself open TO the divine gaze and to internalizethe divine im.lge.

This is 10 exam inc the quesrioll from only one angle though; cvcn in India,it takes an a([ of will on the part of the worshipper to worship the god, andthe worshipper is also an agent with respcct to the one being worshipped. Thereception (as a patient) of dar.lhall from the god is contingent on the transitiveaction of 't'lking' (,Iarslul/l Ima) initiated b~ the recipient. Eck cites StellaKramrisch's account of'seeing' as transitive form of agency:

Seeing ... is a going fonh of the sight towards the obj«t. Sight lOuehcs it amI acquin:sits form. ""oueh is the ultimate connection by which the' isible yields to being grasped.While the e:ye: lOuchcs the object Ihe \ ilalit~, that pulsates in it is oommunicued ..(Kr.1mrisch 19;6: (36)

These remarks ofKramrisch are informed by her knowledge of Sanskrit philo-­sophical wrilings. Ancient Indian philosophers held views similar to those ofthe Platonist.s, that sight was an 'cxtromissive' sense, the eye sending alii invis­ible beams or rays through the air, which touched the objects of sight at theirsurfaces. One philosopher, Caraka, argued that indccd we only have onc sensc,the scnse of touch, of which sight, hearing, etc. are just more subtle forms (Sinha[934.). Other philosophers dis:lgreetl, but the consenSllS was none the less thatseeing was, like touching, a form of contact. Epieurus' theory (see above), bycontrast, is 'intromissi\'e'; the eidola emanate from the object and enter the eye.But Lucretius shows that seeing was, e\'en so, equated with touching by theEpicureans. The (gossamer, but physical) idols were touched at the surface ofthe eye, not at their own surfaces, \·ia the eye-beams. I do not think that theEpicurean intromissi'T theory of \'ision by 'idols' had an Indian counterpart, soone could nOl anribute the pre\'alencc of image-worship to its influence. Out thealternative extromissi\'e theory is, if anything, e\'en more explanatory, becauseit forges :l direo link bet\n:cn 'seeing' (Jarsh(fII) and other types of physicalinteraction with thc image, such liS touching, anointing, and so on. These tactileforms of homage are very important clements in Hindu image-worship.

Certain Indian philosophers (Sinha 11J34) made the analogy between seeingand the use of'l stick by the blind, in order to ascenain the shape of ex lerna Iobjects. This materialistic conception of seeing is l'ellected in Kramrisch'sstatement about darshall; seeing crcates a physical bridge between one beingand another. I-lim's basic insight rhat in relation to 'images' there really isno distinction bel\\een 'similarity' and 'contact' is fully brought into the openhere, without the ncccl to invoke the Epicurean parallel.

Darshun is thus very much of a 1\\o--way affair. The gaze directed by thegod towards the worshipper confers his blessing; com·ersely, the worshipperreaches Ollt and touches the god. The result is union with the god, a mergingof consciollsnessl.'S according to the de\-otionalisl inlcrprelation. This bringsback the issue ofreciprociry and intersubjectivity in the relationship betweenthe image (the index) and the recipient. It is clearly germane to the general the­sis argued in these chapters that intersubjcctivity between persons and indexes,particularly indexes which, like images of the gods, arc human in form, shOll leibe possible. It Cllnnot 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, .llld that the god/dcvotce relationship is II social one, absolutelycomparable to the relationship bctwc<:n the devotee and allother human per­SOil. I-Iowc\cr, it is tOO easy ju~t to accept this as an ethnographic, descriptive,

lIS 1Jtr D;s/ribllud Pamn 1'h1' Di$lrihlUtd {'rrSOIl 119

fact, without a (It.-eper questioning of rhe cogniti\'e basis of this relationship.For all that the devotee asserts, and truly believes, thaI a union of minds isachie\'ed between mortal devotee and immortal divinity, the dCVOIl.X· none theless aL'>() hcs in a world of ordinar} objects, mere things dcwid of imputedsubjectivity, in which the distinction between human bcings-posse!\...cd ofhuman-being-like consciousness and agency-and inert 'things' is readilydrawn. The devotcc docs 1"IOW that the image of the god is only an image, nOI

made of flesh and blood; and if, perchance, the image moves, or speaks, orseems to drink milk, that is a miracle, a rcmarbblc occurrent:c because sounexpected. Devotion is enhanced hy such manifestations, but it is admitTedthat the gods only vouchsafe miracles when faith in them has reached a lowebb, and needs to be buttressed by extraordinary h'lppcnings. True devotion isattainable, idC'ally, without mirncles of this or any Olhcr kind.

The animacy and imputed subjeClivilY of the idol is nor attained except bysumlOunting lhe Slark difference be[\\"Cen an inerl image and a living being.How does lhis happen?

So far as the Hindu material is concerned, the key [0 lhe process of anima­lion seems, initially at leasl, to depend on the logic of looking and being seen.Imagislic devotion is a \'isual act (as opposed to prayer, ClC.) and il is accom­plished entirely by looking. Specifically, it is accomplished by looking into theeyes of the god; union comes from eye contact, not the sludy of all the otherdetails lhe image may show, which indil.:ate the identity and attribUlc.<; of thegod and which add to the general effect without being devotionally L'Ssenrial.The eyes of the god, which gaze ,It the devotee, mirror the action of the devo­tee, who gazes at the god. Sometimes (as in Jain temples) the eyes of images arcset with lillie mirrors, so that the devotee can sec himself or herself reflectedin the imagc's eye in the act of looking. Even in the absence of actual minors,the image, so far as its ocular acti\'ity goes, reflect<; the action of the devolee(F;g.7-7/').

Animacy takes iLS origin from this ocular exchange, becaUse, e\en if one docsnot take a mystical attitude towards images, one is none the less entitled to

apply action \'erh::; like '1001.:' (or 'smile', 'gesticulale', etc.) to them. A perfectsceptiecan S.:lY, in fact is obliged to S::l), lhat an idol 'looks' in a particular dirL'C­tion; the remark would pass unnoticed because e\'erybody acccpLo; that the cri­terion for idols 'looking' is that their eyes should be open and pointed in aparticular dircction. The question is, what do idols scc whell they look? Whatthe devotee sees is the idol looking at him or her, performing an act oflooking,mirroring his or her o\\'n. It is not mysticism on rhe dcvotee's part whichresults in the practical inference rh.1l rhe image 'sees' the devotee, because weonly ever know what Olher persons arc seeing by knowing what they arc look­ing at. The sceplic \muld say 'the idol is blind-il Cannot see anydlillg', butncn so, to be blind is to be unable to see whnl one looks at, which hardlyb:mishe.o; the residual animac~ of images, since a disabilily implil.."S a pOlcnlial

1'1(;. /.7/1. ~~nloshi .\b. Colour lilhO(l:rJph. R G. 'shanna 5£1/1,,'(0 Shanna P,ClUre l'ubliCJlio"-~.

Bomba~.L1960

120 Till' Vis/rih,,/,,} Persoll n,t l)islrihull'd PI'WHI 121

ability. And the devotee need nut draw this inference, in filet, cannot, becllllscthe situation is defmed in terms of the devotee's own agency and result; thedevotee looks and sees. The imagc-as-mirror is doing what the devotee isdoing, therefore, the image also looks and sees. One docs nOf even logicall~

have fO impute 'life' ro an image to assert that an image <.-an see; after all, peo­ple often Spt::.J.\.: of cameras as 'seeing' things, withoul implying that camerashave life.

But the inference that if idols L'3n look, they can see, is not drawn explicitly;if il were, it might be open to sceptical objt-nioos. It is prOTecTed from judge­ment by a funher consequence of fhe mirror-efTect. This is the logical regres­sion set up by seeing and being seen. Eye-contact, nHltuallooking, is a basicmechanism for intersubjeetivity because to look into another's eyes is not juStto see the other, but to see the othel" seeing you (I sec you see me sec you secme etc.). Eye-contact prompts self-;\warencss of how one appears to the other,at which poim (Jnc sees oneself 'from the outside' as if onc wcre, oneself, anobject (or an idol). EJc-contaet seems to give direct access to other mindsbecause the subjeCT s(:cs herself as an object, from the point of \'iew of the otheras a subject. E}·c-contaet is the basic modality of'second-arder intentionality',awareness of thc other (person) as an intentional subj(:ct. Thus, in imag(:­worship, the de'·otce does not just see the idol, but St.'l.'S herself (as an object)being seen by the idol (as a subject). The idol's 'seeing' is built into the devotee'sown self-a\\areness at one remO\·c as the object which is S(.'Cn by the idol. Shesees herself as the idol sees her, kneeling before it, gazing upwards. In that shecan see herself seeing the idol, the idol must sec her, because when she sees her­self seeing the idol (from her point of vie\\, a datum of immediate experience)rhe idol is seen hy her as seeing her. The 'idol seeing her' is a nested (Oll/pOllellt

of 'her seeing herself seeing the idol'. The net result of lhe regression wherebydevotee's :md idol's perspecti\'cs hecome logically intcnligitaTed ,,·irh one anotherin rhis way is a kind of optical oseilhuion in which idol's ami devotee's per­spectives ;hifl bad and forth with such rapidity that interpersonal boundariesare elTaced and 'union' is achie\ed. This mutual encompassment \"i:1 e~e­

contact is represented diagrammatically in Fig. 7.7/2. I shall return to the

Keclpienl-AD,,",,~

~Index-psees ~ Idol

~Index-A _ Reclpicnt-P

Idol ~ lJe\'Olee,-A-..

Rcclpicnt·A lndex-PDevotee - Idol

see5~

etc. ..

FIG·7·7!Z. Ocular eschaul;cas the medium for;111 cr~uhje(;li\'i{ies

significant role of the 'eJcs' in idol-worship, below (Sect. 7.12) in connectionwith the consecration pnx:edures for images.

7.8. AT/imism and Alllhrupolf/orphism

The worship of images, or aniconie indexes of dh'ine pre."Cnce, such as stoncs,springs, and trtes, has preoccupied amhropologisL" from the beginning. Tylor,in the spirit of his times, sought an explanation in a purportcd dilTerence of'mentalities' between primitivc ami civilized people; primitive people wereanimists, whereas we arc not. Trior imagined that The belief in spirits, and 'thesupernatural' generally, was imphmred in thc human mind through the experi­ence of dre:\lllS, and arose from conceptual confusion, bur this was merely aself-serving supposition on his part. In effect, thl: theory of animism is merelya dassifiC:llory do·ice, which, like the concept of ritual, serves to separate beha­\ iour that we think we (:an underSland and sympathize with, from behaviourwhich seems to us superstitious and perwrse.

Howcver, the Tylorian conccpt of animism can be made into a more ser­\·iccable anal) tical tool if it is abstracted fmm the essenrially pcjorati\·e contextof Victorian positiyistic thought. Guthrie, in a recent study of the founda­tions of religious belief, has argued that 'anthropomorphism'-the tendencyto impute human attributes such as will, agency, and responsiveness to sUI>­posedly 'inanimate' entities-is an abiding fcature of human cognition (Guthrie1993). His cognirive argument is That, strategically, it is always safer 10 imputcthe higheST degree of organization possible (such as animacy) to any givenobject of experience. It is better, he says, to prl:SU1llc that a boulder is a bear(and be wrong) than to presume that a bear is a boulder (and be wrong). I lisargument carries weight, and hc certainly has no difficulty whatsoever in ;Imas­sing copious examples ofanthropomorphism, not just in religious contexts, butin e\'eryday perception and (:ognifion, in the arL.., and c,·en in the sciences. Quiterightly, Guthrie emphasizes Ihat anthropomorphism is not a phenomenonrestrictcd to children, so-called 'animists', or o·cn to adult religious belie\·crs.As a non-religious adult, and a participant in an adyanced technological ci\·il­i7..ation, 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 tharto say that one attributes 'animacy' or 'anthropomorphism' LO something docsnot explain what .1 thing must be or do to count as 'animate' or 'anthropomorphic'.

It is not animism, :mthropomorphism, or anything likc it, to auribute 'life'to a tree, which adults in our society agree is a living thing-though childrenunder:; or so may disagrel: (ClfCl' 1985). On the othef hand, it certainly isanimism 10 attribute the capacit) 10 'hear' prayers to a stone idol, bur thardocs not necessarily imply the belief that the stone or idol is 'ali\·c' in a bio­logical sensc. This is abundantI) e\'idcnt from the furore \\hich results \\hen

122 :rIll: Dismhultd Pt:rsfm Tnt Di$lribll/tf1 Ptmm 123

it is reported, as somelimes happens. that idols aClUally do 'come to life' inthc sense of'manifesting biological acth-ity'. When particular statues bleed, orperspire, or mO\'e aboul, these are 'miradl,."S'. But such happenings would notbe miracles if !.he expt..'Ctation was thaI all idols should behne in this wa);in fact, they arc generally expected nOI 10. Idols rna) be animate \\ ithoUl, inother words, being endowed with animal life or :1I.:li\-iIY. In addition, thereare automala, real or imaginary ones, whkh do really move, speak, ;l11d performvarious human-like iKtions, but these are remarkable, not because they arcalive, but because they are IIUI alive, while maintaining rhe appearance of beingso. IL follows that 'ritual' animacy rmd the possession of 'life' in a biologicalsense arc far from being the same thing.

Cogniti\e psychologists believc thai the distinction ben\'een living things(biological organisms) and non-living things is one to \\hich wc arc innatelysensitive, and this is a finding which I do not need to COnTest. I have no doubt(hat worshippers, who address their prayers to stones, are perfectly capable ofthe category distincrion between animals and plants (living things) ,md non­living things such as stones. The distim.:tion between animacy and inanimacythat we ret/uire here cross-cuts the distinction between the living and non­living as 'natural kinds' (Boyer 1996: 86). The go<l who, ar one moment, ismanifestcd in a non-living thing, such as a stone or a statuc, may be manifestin a living thing also, such as a posscs.,<;cd shaman, or a sacred goat or monkey.The worshippers, whose god appears in these contrasted forms, are perfectlycognizant of the difference ben\'een them. The imputation of'animacy' to non~

living things cannot, as Gu!.hrie seems w suggest, rest on people making cat­egory mistakes aboul whether inanima(e objects such as boulders are reallybiologically living things, such as bears.

7.8. I. Stocks and Stones

h seems (hat we must appeal to a differem auribute from the pos!iCssion ofbiological life to define 'animacy'. Should we say that the object is animatenor because we auributc biological life to it, but subje<:(i\-it}"/intentillnalit~,

which is something quite different? As Boyer notes (tI)96: 92), 'the projec­tion of human ph\ sica I fe-,lImes [onto gods, spiriL", CIC.] in general results in aprojection, tacit (;r othef\\ ise, of intentional psycholog), but a projection ofintentional psychology docs not generally entail a projec(ion of any .otherhuman quality'. Jlow can an entity possess 'intentional psychology', WIthoutbeing biologically alive? In this case, ,I worshipper who addresses prayer~ t.o astone must hclie\'e, somehow, that the stone in question, though not a I1v1t1gthing, sees and hears as docs the worshipper, thinks and reac(s as he doc:- andmoreover has the power to plan and e~e<:ule actions. h seems par;ldo.\Ical toimagine t'hat an admilledly non-living rhing could possess rhese attributes.Hut not rcall), "hen one considers whal hel\~ \\eather philosophers have

madc ofaccounting for 'inten(ionality' e\ en among ordinary, living, brca(hing,human beings.

The question as to the manner in \\hich representational indexes-idol..­can be apprehendcd as sociaillthers, as repositories of agency and sensibility,seems to raise (he question of'apparcmly irrational' beliefs and practices. It issurely irrational, or at 1C<ISt strange, ltl 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 'sttangeness' of theirbehaviour as we are, but they also hold, which we do not, thar the cult of theidol is religiously efficllc1ouS, and "ill result in beneficial consl.XJuences forthemselve... and the masters they serve in their capacity as priests. h is not thatthe priests cannot distinguish between stocks and stone... and persons, rather,they hold that in certain wntexts slocKs and stones possess unusual, occult,properties; of which the religiously uninslrueted would remain ignorant, andthe instructed but sceptical, wTong-headedly incredulous. According to Boyer(II}96), religious ideas, such as the 'intentional psychology' of fctishes andidols, survive and prosper ,1S components of culnm1l systemsj1ls1 beulI/sl: theyarc odd and counter-intuitive.

The t/ucsrion that we need tn consider is the nature of the unusual occultcapacities which idols possess, according to believers. Wha( we need to knowis how idol-worshippers square the cirdc benveen 'what they know'-and wha(we know they know-abour stocks and stones, and what the) know about pet­sons and their capacities as intentional agents. They cannot confuse the two,bur it remain.. possible thaI persons havc attributes which can be also possessedby s(ocks and stones wilhout prejudice to their mugor;(Q{ difference from per­sons. That is 10 S3y 'social agents' Gill be drawn from C'J.lcgories which are asdifferent as chalk and cheese (in fact, rather more different) because 'social:lgeney' is nOl defined in terms of 'basic' bi(llogieal allributes (such as inanimatething vs. incarnate person) bm is relational-it docs not matter, in ascribing'social agent' stams, what a (hing (or a person) 'is' in itself; what l1la[(ers iswhere it stands in a network of social rcla(ions. All Ihat may be ne<:essary forstocks and stones to bt..'(:ome 'social agents' in the sense that we require, is thatthere should be actual human pcrsonsfagenL" 'in the neighbourhood' of theseinert objects, norlhat they should be biologiL'":llly human persons thcmsches.

This is nOI as bizarre a claim as it sounds. We certainly do not ha\'c to pos­mlate a p:lrtieular 'mentality' (primitive, uncri(ical, gullible, etc.) to accountfor idol~try; the worship of images is compatible with :1I1 extreme degree ofphilosophic:ll :lnd crilical acumen, as the example of textual Hinduism :lmplydemonstrates, besides numerous treatises on 'thcurgy' (the creation of gods)from classical antiquity (e.g. Proclus). Rather Ihan approach the problem froman initi~1 assumption of the essential stupidity of idol-worshippers, we shouldremind oursehes of the difficulties \\ hich assail (Yer~ de\'er) philosophers, when(hey scek to account for the 'agenc~-', not of stocks and stones, but of human

124 The Di.>lri/mll'tl Pawn Th.. Disirillll!<,t! Persoll 125

hcillgJ> thcmsclves. If (Western) philosophers havc a hard timc pin-pointingcxactly what makes the difference between human persons engaging in ',letions'and mere things obcyingL'ausallaws then we arc in a better position to undersL1ndwhy some pL'Ople appear to be (contextually) indifferenllo lhis dislinction.

The f.'tCI is Ihal il remains a controversial philosophical problem 10 distin­guish belween 'actions' (slemming from agency) and 'happenings' resultingfrom material Lllu.sation. What is the basis of the dislinction we fL'CI we mUSldraw berween a person performing lhe action of raising their arm (intention­ally) and lhe Same physical movemem occurring involuntarily, say, as theresult of some malfunctioning of rhe autonomous nervous sysrem? In olherwords, rhere is a sense in which human beings are rhemselves 'slods andstones'-only rather twitchy ones-and when human beings arc ask-cp, insen­sihle, or of course, dead, the resemblance becomes much closer. The 'philo­sophy of action' is devoted to dnising and testing crileria for justifying thedistinction that we intuitively make between our capacit)' to behave as per­son/agents and our simuhancous capacity 10 be, and beha ve as, Lhings, or mere'crearures' uncndowed with full human agency.

I do not need to canHss, still less evaluate, all the theses \\hich differentphilosophers ha\·e ad\·anccd in the search for a solution to this problem. ~ly

point is JUSt that philosophers do detect a serious problem here; from which Ithink that we can reasonably infer that it is in fact awkward to diffcrentiateberween wh:u people do 'as persons' and what people do 'as things', And ifphilosophers (who know perfectly \\ell that in [he relevant senses persons andthings are different) cannot always agree about 'just how' they differ, then theim'erse proposition also holds; if pressed, we are not rcally vcr)' sure abour 'justhow' an idol is /lOf a person-evcn though we are perfectly certain it isn't.We C'dnnOl rcly on simple arguments like this: 'I am a person, 1 have a beatinghearl, a tcmpcrature of 98... degrees, etc. whereas this statue is stonc-coldand has no heartbeat, ergo, it is not a person'. We could heat rhe starue lO98-4 degrecs, givc it a heart, etc. but it still would not qualify because ir wouldonly possess, after lhese modificnions, attributes which human beings pos­sess 'as things', not as pcrsonsfagents. Yet such arc the arguments which arcadvanecd 3g'dinst idolatry by sceplics:

Lila Lajpat Rai de~erihes how the founder lof the Arya Samaj, a 'reformed' Hindu sect,opposed to rhe use of imagesl first got his insight into rhc wrongness of iclobtr)'. Hcwas sct, as a lad of fourtecn, to watch an image of the god Shiva, in a H:mplc at night.I Ie saw :1 mouse run ovcr the god's body and the god rcmain motionless. The shockeOllvinced him, Lajp:ll R:I; wrote, that 'the image could not be Shiva himself, ,IS wastaught by the prieSThood'. (Bcvan 1940: 34)

Such an argument could have been convincing only to one who had already(no doubt as a result of Christian-Protestant ascendancy in British India)decided that idol-worship was backward and futilc. Shiva's indifference 10 thc

mouse could equally well have been interpreted as his superhuman asceticismmanifesting itself-indeed eompllrahle feats have been attempted by asceticswhose humanity is not in doubt. Some, for instance, arc reputed to haveinduced birds to nL'Sr on their motionless bodies. Any such argument can cutbOlh ways: if idols are nm whal they pretend to be, or arc pretended to be, itis not because they arc 'things'. Human beings arc also 'things'. If an effigypossessed every single thing·auribute of a human being, it could still be 'juslan effigy' and be unwotlhy of wOI'ship; conversely, an effigy could possessno identifiable lhing-arrribule of a human being and be worthy of devotionne\'erlhclcss.

Suppose we perfoml Ihe thoughl experiment of gradually enriching an idolwirh more and more of the all'ribulCS of a 'gcnuine' living being, Temperature,heartbeaT, mobility, the abiliry to utler words, ro play tennis, to ... well, youname it. Does the idol become more worship-able as a result? By no mcans:either, by this procedure, the idol becomes a common-or-garden human per­son, whom it would be senseless to worship, or the idol remains an idol buttakes on the status of :l.I1 automaton of extraordinary yerisimilitude worth\'of exhibition and admiration in an eslablishment like' Disneyland or ~hdam~Tussaud's, but nor of reverence or de\otion. The criticism of idolatry on theground.. that idols are nOI 'ali\'e' as human beings are (biologically) alive, orthat idols are not realistic automata, bUI only statues, misses the point on bOthcounts. The idol is worshipped becluse il is neilher a person, nor a miraculousmachine, but a god,

Whate\'cr the atlribules possessed by idols which render [hem religiouslyefficacious as a locus for person-to-pcrson encounlers with divinities, [heseauributes cannot be confirmed or disproved by physical tests such as the pres­ence of a pulse, respiration, ingestion and elimination, the ability lO move orspeak, a natural distaste for mice, and so on. None of these attributes figures inphilosophical aucmplS to dislinguish octwL'Cn pcrS()Il/agcn[s and mere things,machines, effigies, illusions, and so on. Currently, many philosophers agreethat 'agency' implies the possession of a mind which 'intends' actions prior toperforming them. 'Not mO\'ing' is an 'al;tion' in this sense. 'Shiva (the god)did not move because he intended to stay still' is a perfectly reasonable inter­pretation for the sccne wit'nessed by Day.manda, on the assumption that Shivawas, as [he priests averrcd, present as a person!'lgenr, in his image. He mighthavc had various reasons for reft',lining from action; first (as just mentioned),becausc Shiva is thc prototypical immobile ascetic, and secondly, bee.lllseShiva (who created thc whole world) imbucd stonc, the material of his visible'hod}" in [his inst,l11ec, with the property of absolute rigidity; Shiva lVasobserving the 'rules' for slOnc objects, such as idols, whicb were of his ownmaking, ul[imarciy. Hindu theology, moreover, pustulates [hat gods, such asShiva, voluntarily 'sauifice' their freedom of movcment, imprisoning them­seh'es in stone idols for the bcnefit of devotees.

126 TlJt DiSlrihllt~tI Pawn 71'1' Distrlbuud PawlI 12i

This is the b'l'calcst grace of the l.uro, that being free He becomes bound, being inde­pendem He bt.°(,:omcs dependent for all l1is service on His OC\'olce ... In other fonnsthe man lxlongcd 10 God hUE behold the supreme S3.I.-Tifi«: of Is\"ara LVishnuJ here theAlmighty becomes the property of the dcmltt .. _He Clrries Him aoom, fans Ilim,fetds Him, pla~s \\ith Him-yea, the Infinite has become finite. that the child soul rna)b'TaSP, undcl'SLanJ and love I lim. (rillai l.ohcarya, cited in Eel: 1985: 46, diacritil:Srcmo\'cd)

There arc no a priori logico-philosophical grounds for rejecting the thesisthat Shiv:! 'intended to stand still, and thus stood still, when the mouse ranover him' (ami was an agent in so doing). Because there are no 'material tests'for the possession, or non-possession, of ngem:y, rhere is nothing to preventus from asserting, if we wish to, that the behaviour of a statue (standingstill) occurs because the Sl:l.llle has a mind, intends to stand still, and does as aconsequence of this prior intention stand still.

7.9. External and intemal CQnceptions ofAgency

How, in practkc, do we attribute 'agency', 'intentional ps~chology'-the pos­session of a mind, consciousness, etc. to 'social mhers'? If we knew the answerto this question we might be in a better position to define precisely what sub­set ofhwnan-like attributes idols as 'social others' are belie\'cd to possess, giventrun they are probably nor (on Boyer's argument) bc1ie\·ed to possess all ofthese attributes. I rhink it is fair to say that our attribution of 'intentional psy­chology' to anything (a person, an animnl, a computer, :1 car, or a stone idol)has two aspects to it, which:ll first glance seem to be rather distinct. The firstof these one could call the 'external :lspect' or the 'practical' aspect of agency­attribution. According to Wittgenstein, and a great many other subsequentphilosophers, the possession of a mind is something we attribute to others,provisionally, on the basis of our intuition that their behaviour (e.g. theirlinguistic behaviour) foUows somc 'rule' which, in principle, we may r(.'<.:on­struct (Winch 1958). If I can get along with the other in the give-and-take ofinteraction, if our practical efforts to deal with one another work out, then theother is a producer of intelligible (meaningful) beha"iour, and hence has amind, intentions, volitions, ctc. I Lilnnot rcally tell, from the outside, whetherthe 'other' is a zombie or an automaton, who/which mimics the behaviour ofan ordinary human being but docs not haye any of the 'inner experienu:s' wehabitually assod:lte with this behaviour. But this docs l10t matter becausethe whole pannply of ,mind' is not" a series orinner, private experiences at all,but is OUI' there, in the public dom'lin, as language, pmctices, romines, rulesof the game, etc.; that is, 'forms of life'. C.'111 this the 'extcrnalisr' theory ofagency-at tribu Iion.

However compelling in the hands of beha\·iouristic.lllr inclined philoso­phers, the 'exlernalisl' Iheory of agency has a weakness, namely, that il docs

nor immediately square with psychological ill\'Lostigations of the wa} in whichchildren and ordinary people (not philosophers) seem to approach the sameproblem. It seems that ordinar} human beings arc 'natural dualists', inclined,more or less from day one, (0 belieye in some kind of 'ghost in the machine'and to attribute the beha\ iour of social others to the m~lIlal r~presenla';o"sdustollirrs hau ';11 flit;r hraJs'. Bcha\'iour is caused by factors which well up fromwithin the person, thoughts, wishes, intentions, etc. Minds are hidden awa\'inside people, rather than being manifested in bctwL'Cn them, in the pUbli~space in which inlcraction r:lkes place, as the externalist theory seems to bc say­ing. The 'internalist' theory of mind, according to Lhe cognitive psychologists,is:1 'modulc'-a kind oftheo!'y (or principle ofintcrprel'ation) with which weare born, along with the principle that there is a basic distinction betweenli\'ing things and non-1i\<ing thil1b"S' When Boyer speaks of'intentional psycho­IOg)·' he is referring to this (pumtivcly innate) module.

~'lany philosophers bc1ie\'e that the notion th:lt genuine 'persons' are beingswhose beha\ iour is caused by the mental representations that they ha\·c 'intheir heads' is not just a truth of common sense, hut is pcrfectl~ dcfensiblephilosophically, if formulated with due carc (e.g. Fodor 1(94). However, \\cdo not need to consider the argument... for and against the 'mentalist' (i.e. inter­nalist) position in the philosophy of mind in any detail. All that we need to beawarc of is that these two rout,es towards 'agency'-auributions exist. Let liS

consider further somc of Ihe differences between thelTl. For a start, they eachbegin with a rather different problem-dcfinition. The externalist theory is notreally about the 'psyche' or 'consciousness'; it is an ,lCcount of inLcrsubjcct'iv­it}' rather than subjectivity, and it explains how it is that intersubjectivityis quite possible, even in the absence of some telepathic lTleans of enteringanother person's skull and having his thoughts, feeling his pains, and so on.Uecause the extcrnalist theory is about inrersubjecti\ ity, it is popul:lr with soci­ologists, many of whom are much more beha\"iouristic in their thinking thanthey rcalize or acknowledge. To cite a case in point, the leading social theoristof today, Pierre Bourdieu, acknowledges a debt to both Wittgenstein and to the(subtle) bchayiouristlearning-theorist, Hull. Bourdieu's im·aluable concept ofthe 'habitus'-the sedimented residue of past social interaction which struc­tures ongoing interaction-is not a transcription of common-sense mcntalismor 'folk psychology', but is precisely a notion ofl1lind externalized in routine,practices, that is, the prevailing 'form of life'. Sociologists h3\'e 10 be 'exrcr­nalisLs' becltlse culture and soci:ll institutions :H'C eXlernal, interactive, pro­cessual, historical realities, not states of mind. Sociologists cannot he 'pure'mentalists bec:HJse, apart from anything else, thcy are concerned with :Ictionsin the light of I'heir (OI/i"eqllt:llres, and we arc all roo lI"eli aware [hat our actionsrarely if e'·er have precisely lhe consequences \\c hoped or expecled them tohave. 50;1 theory which only relates aClions to (inner, prior) intentions, evenif adequale ps~chologieally. is ~iologic;dly inadequate.

128 The Di.{/rilmfa/ Person n,e f)iS/ril/lll('d {'emm 129

Since this is a I'rc;l1isc on lhe sociology of art, there afC good reasons why I,in particular, should pay attention to the cxrernalist theory of agency. Thesimplest solution 10 the problem of idolaTry is an 'external' one on tbe follow­ing lines: idols arc 'social others' to the extent that, and because, rhey obeythe social rules laid dowll for idols as co-prcscnr others (gods) in idol-form.'rhus, according to Eck (1985: 48), a Hindu deity in idol-form is essentially anhonoured 'guest' 10 whom the dnolce pays homage in the form of nurture(offering food, fanning the flies away, ClC.). The appropriate behaviour for'guests' of e.xtremcly high status, is, in faer, ro do more or less exaclly wharidols do: accept what is offered with impenurbable dignil)' and impassivity.1dols 'produce intelligible beha\'lour' which conforms to cerrain expectations.

Of course, idols do not apparently 'do' anything; they gcncrally just standLhcrc, being immobile. 'I'his seems like an odd form of'intelligible' behal'iour,but it is not; ,he Life Guards outsidc Whitehall barracks produce exactly thistype of'intelligihle' immobility and apparent insensibility as a beha\'iour, andthey arc quinressemially 'social others' while they do so. When c\·il-mindedtourists poke umbrellas at their horses, they do not suddenly behave 'out ofcharacter' and curse their lOrmentors: they preserve their icy indifferem.'c.They are playing the game, and it is a game we em readily participate in. Theyare like the Scpik warriors whom Harrison describes as aspiring, in totemicritual performances, (0 emulate the impervious spirirnalit)' of totcmic sacra,the ceremonial effigies of spirits which line the intcrior of their long-house(Harrison 1983.' 118).

This 'Dionysi:m' aggrJndisement of tht: self, reserved for a selecl few, is I'iewed not asa l:dt:hl":tlion of subjecli\ ity but as the rc\-ersc: the depersonalisation of the actors intohuman equivalents of ritual objects, like the masks, staluan' and otht:r satT:1 whichfigure in Ihe men's cult as embodimt:l1tS of ritual potency. "ien wntCX"lUally suspendtht:ir 'normal' social idenlitit:S but gain power, in the foml ofthe impael which highlr­charged symbolic objt'Cts have on the subjecti\·jtics of others.

Howe\'Cr, I agree thaI such ~m extemalist interpretation of agency-anriblltionto idols seems too simple by half. Although the Lifc Guards may preserve theirstony dcmeanOlll'S when their horscs get poked, none the less, we know, andthe)" know we know, thar they are silently thinking 'buggel' oIT' or words to[hat effect. And that is why their se1f:'conrrol is so remarkable and admir­able. It is all vcry well to say that an idol which stands immobile is producingintelligible idol-bchaviour according to the accepted social rules, LHJt unlessthere is something going on inside the idol which corresponds to the 'innerlife, unspoken thoughts, etc.' whieh we alt'ribute to r .ife Guards, or Sepikwarriors, then docs this behaviour 'count'? There arc twu ansllers to this kinduf objection. rirst of all, although idols may not produce much \'isiblc beha­viour, they may none the less be \"CI"Y 'active' il1\'isibly. that is, most of theiractions t:lke place 'off-stage' as it \\"Ct"C. They can be making the crops grow,

conf(JUnding traitors and plolters, or keeping the sun alighr, or enjoying them_selvcs in heaven, erc. And these ~lct'ions (except perhaps the last) have visibleconsequences 'elsewhere', though not in :my change in the bodily demeanourof rhe idol per sc.

Secondly, an idol who due.... nOf respond actively (hy moving or speaking) isnone the less 'active' as a patient wirh respect to the agency ofothers. And thismar be enough. Children's pia) with dolls C'Jn sene to illustrate Ihis type ofpassi\c agency. Dolls, cI'en those which open and dose their e~es, emil cries,or even wet rhemselves, never produce any behaviour which is nOI dirccrlvunder lhe control of thc nurture-pro\ iding play-mother. 'fhe playing childknows this perfectly \\'CIl, but that docs noT prewnt them havlng the livelicslsensation that the doll is an alter ego and a significant social othcr. 0011 play isso totally !>atisfying just because of this passi\'ity; the doll docs just whateverthe child wants; submits (0 undressing and getting dressed again, sleeping,waking, and eating, and CI'en, if m.x.'t..~<;ary, being 'naughty' when the child is inthe mood to inflict a smacking. The doll's 'thoughts' and inner life (which cer­tainly are attributed to it Ilhile pia) is in progress) are a reflex of the child'sown thoughts, which include Ihe doll as a pa.......ive being whose thinking is done'for it' just as dressing and undressing have to be done for it as \vcll. The play­ing child thinks for her doll as \Iell as doing e\'erything else. We could inter­pret the thought-proccs.'OCS of idols along the same lines, that is, as somethingthat deHm.:cs do for idols, which t'3n nunc t~e It........ be attributed (in context) to

them. !:Iee-dUse idols (like dolls) arc wholly 'passive' others, they exhibit 'pas­sive agency', the kind of agency atrributdble ro social others who or which, bydefinition, are onl} the largel of agenC) , ne\'er the independent source. (Theargument of this paragraph relate.... to the analysis of Hindu dan/wIl, abo\'e,Sect. 7.7.)

Ilowever, I agree that neither of these responses quite measures up to theobjection that idols, because of their behavioural ineffectuality, cannot be con­sidered 'agents' in the full sense. There does seem to be a basic differenceberween the idol-like Life GU;lr<! whose mind seethes with unuttered curses,and the genuine idol whose dignified gaze betokens, for us, no such inner life.Whoever imagines th;ll the idol is conscious, thinking, intentional, erc, isattributing 'mental slatcs' to lhe idol which have implications, not just forthe external relations between lhe idol and the devotee (and the form of lifein which rhey co-pal'licip'lte), but for the 'inner structure' of the idol, thatis, that it has somerhing iI/side it 'which thinks' or 'with which it thinks'. Theidol may not be biologiCill1y a 'living' rhing' but, if it has 'intentional psycho­logy' attribllled 10 it, then il has something like a spirit, a soul, an ego, lodgedwithin it.

This is certainly true, elhnographic'llly and psychologically, because of theinnateness of the 'theory-of-mind module' which attributes intentionality topersons (and things as wdl, under eert:tin circumstances) as a component of

130 The Distrihuted PerSOIl lhl" Distrlbu/I'd Person 131

what Schutz calls 'the natural attitude'. The problems which assail spontaneousmentalism arc not to do with demonstrating its existence, but with pursuing itsimplications. Let us say that 'intenrional psychology' (that which is attributedto idols according to the 'internalist' approach to agency) consists of somethinglike a 'conscious self' as experienced in the 'first person singular'. This is 'theMind's" I'" (Hofstadter and Dennett 198z). 'rhe trouble with this mysterious'I' is not that anybody truly disbelieves in it, bur that nOlhing in the world, nophysically identifiable thing, really seems TO correspond to it. The cxternalisttheon' of mind does not seem to give one anv real reason for believing in thisentit): in which we all do, none the less, believe: and on the basis of such abelief~ interpret and predict the social behaviour of others. The philosophicalproblem of mentalism (of intentionalism generally) is succinCTly summed up byDennett in the following passage:

Fin/, the only psychology that could possibly succeed in explaining the complexitiesof human activity must posit internal repres~ntations. This premise has been deemedobvious by just about everyone eXl;ept the radical behaviorists ... Descartes doubtedalmost en~rything but this. for the British Empiril;ists, the internal representationswere called id~as, sensations, impressions; more recently psychologisTS have talked ofhypotheses, maps, schemas, images, propositions, engrams, neural signals, even holo­grams and whole innate theories. So the first premise is quite invulnerable, or at anyrate it has an impressive mandate ... But, suond, nothing is intrinsil;ally a rc res~nt:J.­

tion of anyt.!!0g;_SO)llcrhm.g_ i.:.a repre nta Ion on y or or /0 ~Il1eon~;~epres­

entatIon or svstem of reprcsentations thus requires at least one user or il/ferpreler of." the represen;tion wht) is extcrnal to it. Any sUl;h inLerpreter must hayc:l ~ariety oT?1 <':'psychologieal or intcntional traits ... iT must bc capable of a variety of tOmpreilcl/siol/,

and must h.we bc1ief~ and goals (so it can lise rhe reprcsentation LO inform itself and thusassist it in reaching its goals). Such an interpreter is then a sort ofholllUIll;Ulus.

Therefore, psychology wi/hflu/ homunculi is impossible. But psychology ullth

homunculi is doomed to circularity or infinite rcgress, so psychology is impossible.((Dennett 1979: 119-.;1'2) ,_ ' 1-- \ --I,,,,'--- \ c\.o> \ ) V~, ~ 'f' "-. .

Dennett argues that this problem is surmountable, not by gettmg nd ofhomunculi but bv ha~'ing lots and lots of them; 'stupid' single-task homunculidoing low-'lcvel t;sks and relaying the results to more intelligent homunculidoing higher-level processing tasks. There is nor anyone 'mind' but a pande­monium of homunculi gcnerating representations and selecting among thosethat have been generated those which arc of usc in fulfilling rhe organism'sneeds. My purpose in citing Dennett on homunculi, however, is not to inrr(}­duce his brilliant, artificial-intelligence-based theory of mind, which seeks tobreak the deadlock between the various compelling rcasons to accept somemodified form of beha\'iouris111, and the need TO explain consciousness as weactnall\' experience it as psychological suhjects. Dcnnettt, like all modernphilos~phersof mind, is writing about real human beings, not idols. The pointthaI interests me is that Dcnnett is suggesting thrtl in so far as we concei\'C

of human beings as intentional agents because they generate and respond tomental representations, then we arc obliged £0 'split' them (internally) intotwo; the one who 'has' the representations (pcrccptions, ideas, etc.) and the onewho interprets thcm (see Dennett, ibid., ch. 5). \\lhat I derive from this is thecognitive naturalness of the idea of the mind (or soul, spirit, etc.) as a homun­culus; that is, like a person but (onta/ned /pithin a person. That is to say apredictable conscquence ofour (possibly innate) propensity to anribute 'in~:n­tioml psychology' to humans, animals, etc. is attributing a homuncufus-like jiJrmto this 'interpreter' lodged within the other, when the other is being attributcdwith an intcntional psychology. Tha[ is to say, if we are to attempt to 'depict'thc physical realization of thc 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 inncr person who interprets the other's repres­entations) and lodge that homunculus inside the other's body.

Let us return to the idol. We havc established, I hope, [hat the idol is accept­able as a social other on the basis of 'fitting in' to the role expectations for idolsrtS a particular category of social agents, [hat is, primarily passive agents oragcnts whose agency is exercised 'oft:'stage'. Practically and physically, this isperfectly manageable; we just have to stipulate what The idol, to conform to itsrole e.xpectations, shall look like, and manufacture an artefact which has thestipulated external characteristics. But what about' making an idol which, onthe basis of its actual physical characteristics, motivates the attribution TO it ofan intentional psychology? How might we do that?

\Vell, we might not want to: according to Boyer (I996) the attribution ofintentional psychology (or other occult aHribmes) to non-living things is apotent religious idea precisely because it so markedly contradicts two basicassumptions about reality, (i) that living and non-living things arc totallydistinct, and that (ii) intention:ll psychology can only be attributed to livingthings. On this theory, there would be a strong (basically innate) cognitiveprcference for religious objects (attributed with intentional psychology) to beaniconic in form; lhe more bla[antly the supposedly animate ritual object failedto measure up to the normal criteria for animacy, the more enthusiasticallvbelievers would worship it. However, this prediction is not borne out in pra~­tice; where the technical system of rt particular rcligious community includestechniques for manufacturing iconic or anthropomorphic images, idols, etc.,such idols are very often manufactured. J\;loreover, supposedly aniconic reli­gious objects are often locally interpreted in 'iconic' ways. G;)ouman's well­known jibe against the notion of ,realism' in art (J976) c~rtainly applies here,namcly, thaI since everything 'resembles' everything else in at least sOllierespects, everything can, under some interpretation, be regarded as 'depicting'anything you like. Consequcntly an uncarvcd stone can be an iconic repres­entation of a god just as wcll as a minutely carved stone idol which looks muchmore 'realistic' to us.

•

1.l2 Tnt DislrlbllJrtl Pawn 'I'll( t)jslrihllttll J'aSOf/ 133

!-Iowe\·er, I think that we can quite e2sily distinguish between idols in \\ hichiconism, the impulse to depict resemblance, is thematic, and those in II hichiconism is non-thematic, as in litholatr}, pure and simple. So I rephrase thequestion; given that we are dealing with that class of idols in which iconicdepiction of the object of veneration is thematic, how arc we TO indic.lte, byIlleans of piclClrial or sculptural 'mimesis', that the artefac< is endowed with anintentional psychology? There is a simple answer to this: we cannot. There isnothing physical that \\e can imitate here, there is no mimi in objective formthat we can copy and insert into the appropriate place in the idol. 1\0 manerh011 realistically we imitate the outward appearance of the bod}, we fall shortof depicting the soul which, howner, we are detennined to imirate in somef.1shion.

To sav that there is no ideal solution is not to sal' that there are no half­measure~. Even if we cannot depict the mind, \Ie can ;t least depicl lhe possib­ility th:ll there is a mind we cannot depkt. Ry way ofa thought experiment, letus postulate an 'ideal anieonie idol'--,---.;} sphere of perfectly homogeneous mater­ial, actually black basalt. We may suppOSe that the spherical swne idol has amind, intentions, sensibilities, etc. but there is nothing about the material char­act.eristics of the sphere, as such, which aniculates with these beliefs, which areentirely theological and abstract. But leI us modify the spherical idol somewhatby drilling a hole in it, or maybe two holes, which would then probably be seenas ·eyes'. Once lhe sphere was equipped with 'orifi<.:es' of this kind it would bepossible, not just to imagine, abstractly, that it had a mind, perceptions, inten­tions, etc. bur to attach these imaginings to the formal contl'"Jst between theexterior of the sphere, into which the holes were drilled, and the interior, to\\hich these holes give access. Adding fel1tures which apparently make thesphere more 'amhropomOl'phic' (by the addition of eyes, a mouth, etc.) do notjust ser\e the purpose of making the sphere a more realistic 'depiction' of ahuman being, they rcnder it morc spirirtJ:ll, more inward, by opening up routes

of({(tess to this inwardness. The 'imernalist' thcnry of agency (in its informalguise as part of everyday thinking) motivates the development of'representa­tional', if not 'realistic' religious images, because the inner versus outer, mind\'crsus hody l."Olltrast prompts rhe developmem of images with 'marked' char­acteristics of inwardness lersus outwardness. PandoxicaIly, the developmentof idols which depil.1: the \·isible, superficial, features of the human body makepossible rhe abduction of the 'invisiblc' mind, a\\arcness, and ,will fro~n thevisible image. The more m:nerially realistic the image, at leasl III certalll keyrespecls, the 1110re spiritl1:l11y it is seen, ..

It would be misleading 10 suppose, though, that the need to arneulate VISU­

ally the contl~st between inner mind and OlUer body k~ds ineluctably tollardsrepresentation31 3rt forms, though this happens. i\ly argument is that theindexical form oflhe mindfbody contnst, is primordiall~ spfltifl1l1mJ (Of/Ulltnc;

the mind is 'internal' enclosed, surrounded, by something «(he bod~) lhat is

non-mind. Now we begin (0 sec why idols :lre so often hoUow em'e1opes,with enclosures, like the hollow ntdllr; stones, or the hollow sorcery-imageswe encountered in Ihe preceding discus.'.ion of Polynesian idolatry (aboye,Sects. 7.5-6).

It is often the case that the human body (with an implied interior indicatedby orifices) is used to index this primordi'll insidc~outside relation. But there3re other ways of achieving this as well. Suppose, instcad of drilling 'eye'holes in the spherical idol, we leaw it. as it is, but pl3ee it in a box, an ark.At this moment ir becomes possible fO rhink nf the spherical idol in a difTerentway; we ca.n easily suppose rhat the stone inside rhe txlX is the locus ofagency,intention, etc. and the ark is the sacred '\·esse!' which, body-like, contains andprotects this locus of agency. Once the idol is in the ark we h:l\e, once more,the physical configuration necessary for thinking of the SLOne ,lS 'opposedlO' something else in the way that the mind (interior) is opposed to the body(exterior). The 'homunculus-efTetl', in other words, I.~n be achiel'cd withoutanthropomorphizing the index, so long as the crucial f~Jture of concentricityand 'containment' is preserved.

7.10. 71te Anima/ion of Idols: The Extema!ist Strategy

There arc thus two b,lSic stratcgics for converting (conceptually) stocks andstones into quasi-persons in artclact-form. The first of these strategies consistsof animating rhe idol by simply stipulating for it a role as a social other. Thesecond consists of pro\'iding it 1\ ith a homunculus, or space for a homunculus,or turning it into a homunculus within some larger entity. I shall discuss the'internal' animation of idols in the nexr section. Here is an example which gi\'CSus a good view of the eXlernalist strategy in action. Conlrary to whal onc mightexpect, the most important im'1ges of the gods of ancient Egypt wcrc not thcmonumclll'al carved figures that h:lve survived to this day, but much smaller,conveniently heftable, idols, as the following description makes clear:

in the lemple of Hathor al DcndeT:lh, there were, among olhers, lhe follo\\ ing sacrt:dstarnes: Ilathor. lX'int«l wood, cojJper, inlaid c~·es, height.i dis, ~ spans, and 2 fing~rs;

Isis. p3inted acacia \Iood. C~'CS inlaid. height I ell; I lotus, paimed "ood, inlaid eyC!o,heighl I ell and I finger. The largcsl, therefore, was scarccl~ of life size; the smallt:!>1onl~ about J6 inches in height. The reason fot this insignific:mcc in size was that forcert.Jin ,Jets of worship lhe images had to be easil) pormhle,

The paltry size and material or thest: little wooden dolls II crt:, however, atoned forby the splendour of lheir ahode, and the rC\crcnec \lith which they were sen'ed. Theshrin~ of the god wa~ in the innermosl chamber of the temple, "hich II as in tOI..ldarkness sale on the enlt\ of the oltlcialing pricsl be:iring artiliciallight. II consistedgeneT:lll~ of a single blocl of SlOne, ont:n, esp(:dall~ in the later periods, uf enormoussize, hel\n infO a house I\hich surrounded \\ilh impenetrable walls lhe image of Ihe

134 The Di,tributed Person The Distributed Person 135

god. The doorway in front was closed with bronze doors, or dOMs 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 tied with a cord ocaring a day seal.Within the shrine, the image of the god repo~ed in a little ark, or portable inner shrinc,which could he lifted out and pIKed upon the barque in which the deity made his jour­neys abroad on stated occasions.

The daily ritual of service to the image was in its main outlines the same in all thetemples ... the procedllre was as follows. Early in the morning the priest of the day,after lustrations, entered the Holy of Holies, bearing incense in a censer, and stoodbefore the shrine. He first loosened the door that closed the shrine, repeating as he didso a stereotyped phrase: 'The cord is broken, and the sealloosened,-l come, I bringthee the eye of Horus [i.c. light, the sunJ' ... As the doors of the shrine opened andthe god W,IS revealed, the priest prosTrated himsclfand chanted 'The !rAtes of heavenare opened, and the nine gods appear radiant, the god .'\J is exalted upon his greatthrone. . Thy beauty belongs to thee, 0 god -"J; thou naked one, clothe thyself.'Taking his vessels, the priest then began to perform the daily toilet of the god. Hesprinkled water on the image twice from four jugs, clothed it with linen wrappingsof white, green, red, and brown, and painted it with green and black paint. Finally, hefed the image, by Ltying before it bread, beef, gee~e, wine, and water, and decoratedits table with Ilowers. (Blaikie 1914: 132)

It is nOl hard to sec thc applicability of the externalist theory to the cult ofthe idols in the temples 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 pointlessunless these life-endowing rituals were literal transpositions of the means inwhich we induce agency in social others in human form, such as children.

Indeed, iL is very hard to rcad 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 hut actually makea cult of, or worship them. 'Piny' behaviour is supposed to take place in con­ceptual brackets, which say 'this is play-so I am not doing what I appear to

be doing' (Bateson HJ36). Children, outside the temple, might piny at beingpriests, and pretend to worship toy gods, bUL this type of make-believe isentirely distinct from the activities of the priests themselves. They were notat play, but at work. They were serious.

Nor is it quite right to say that their actions were 'symbolic'-though ofcourse everything rather depends on how the word 'symbolie' is understood.Offering food to the image of the god is not a pantomime or dumb-show, as ifthere were some alternative 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 offer­ing before it is not symbolic in the sense of 'meaningful', but the 'meaning'

stemmed from the real (causal) outcome of this act offeeding; the god was nolongcr hungry. Jhe3SSe!!£e_of idolatry is [hat it permits real physical inter­

ae/isms to take place between l?ersons and di"villities. To treat such interactionsas 'symbolic' is to miss the point. lmage~ be employed in worship in non­idolatrous ways, as aids to piety but not physical channels of access to thedivinity-the Christian use of rdigious images is supposed to belong to thiscategory, thougll in practice many Christians' use of images is dc/aclo idola­trous, if not admitted to be so. We can only distinguish betweell idolatrousand non-idolatrous usc of religious images because idolatry is in an importantsense not 'symbolic' at all, whereas the use of imagcs as aids to piety, ratherthan physical \'Chic1es of divinity, is symbolic. The Egyptian ceremonies justadduced belong firmly in the category of idolatrous practices, and arc thus real,practical, services performed for divine social others in image-form, not sym~bolic acts.

All the same, the way in which thesc idols were enmeshed in the structuredroutines of daily life can only provide a panial answer to the problem of idol­atry. Therc arc other features of the situation which seem to indicate the oper­ation of different litcrars, which cannot be accounted for in terms of a purely'externalisl' notion of agency. For instance, as noted above, the purely extern­alist theory of agency makes no stipulations as to the physical or visual formof the objcct (index) which is treated as an agent. These Egyptian idols wcrc,in fact, quite realistic representations of the outward appearance of humanpersons. The externalist theory of agency is not in a position to differcntiatebetween 'iconic' idols, such :lS these, and 'aniconic' idols, such as stones, orthe planks of wood, known as baitulia, which the Greeks vmrshipped (:ISAphrodite, Zeus, etc.) hefore, and alongside, their subsequent cult of sculp­tural images of these gods. This indifference towards the iconic properties ofidols is, in a sense, a point in favour of the 'externalist' theory; unless agencywere :1 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. Thc 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 'inrernalist' or homunculus theory. \Vhile it may be true that the agency ofthese idols derived in pan from the way in which lhey were insened into therelational texture of 'cxternal' social praxis and language, this 'passive' agencyis certainly not the whole story. These particular idols werc in fact highlyiconic, and morcO\'er, the description given (by Blaikie, hut deriving directlyli'om Herodotus) emphnsizes particularly [heir 'inlaid eyes'. Tn the next sectionI will describe in derail the consecration ofcontemporary idols, which cruciallyinvolves [he animation of images by providing them with eyes. Eyes arc, of all

idol: temple:: mind: body.

These reflections lead towards a relativiz:llion of the contrast between theextcrnal and internal conceptions of i1gency, sentience, etc. with which I beganthis section. It is ob\'ious that the homunculus, or 'inner person' conception ofngency cssemially reduplicates, within the human person, the relation whichalways exists between a human person and a texturc of external relationships,but within the interior domain, within thc body. This imagcry leads to thc'homunculus within a homunculus' problem which besets this type of theory,according [() its critics. Hut this 'problem' is also an advantage, in that it tendsto blur the distinction between the 'induced' kind of animacy which is imposedexternally on the idol by enmeshing it in praxis, language, social relations andromines, and the 'internal' agency which lhe idol is supposed to possess as a'mind' encapsulated in a surrounding body. Just as the idol, externally, is at thecentre ofa concentric array of relations between persons, so the idol, internally,can he seen as a concentric array of relations betwecn the 'inner' pcrsons-

body orifices, Those which signify 'iIlleriority' (i.e. the posscssion of mind andintentional iTy) most immediately (see the previous discussion of dllf.l"han). '['heparticular atteillion paid to the eyes of these idols arises, not from the need to

represent the body realisticallv, but from the need to represent the bodv insuch a way as to imply that the body is III/I)' a body, and that a much n;oreimportant entity, the mind, is immured within it.

Thus, it is equally important to norc that the animation of these idols wasbeing :lchieved, simlllLmcously, in a qllite different way. The description givenabove emphasizes, hesides the re,llistic form of the idols, their extraordinarvand impressive surroundings. They were kept, except when being served b;'the priests, in a box or ark, which, in turn, was kept in the darkest and mostcentral sanctuary of a vast temple complex, consisting of innumerable lessersanctuaries, shrines, courtyards, barracks and workshops, etc. If we situateourselves, not inside the innermost sanetuarv, but outside in the court van.!with rhe ordinary worshippers (who rarely if ever saw the idols thems~lves)then we may readily imag'ine that the idols (immltred in the temple complex,and animating it like a giant hody) come to stand for 'mind' and interioritynot just by physical resemhlance to the human body, but by becoming theanimating 'minds' of the huge, busy, and awe-inspiring temple complex. Just:lS the 'mind' is conceived of as all interior person, a homunculus, within thebody, so the idols are homunculi within the 'body' of the temple. And it istrue that idols, even very representational idols, :lre invariably presented in asetting, a temple, a shrine or an ark, a sacred space of some kind, which hasthe effect of emphasizing their interiority, their secludedness and (relative)inaccessibility, as wcll as their m:1jesty. The seclusion of the idol has, .IUto­

matical1y, the effect of motivating tile abduction of agency, on the b:1Sis of theequation:

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

137:/'lie Disrritl/lled Perwn

, Curiously though, contemporary Runllans, according to the ethnographer .>\lain Rahadzan have aquite differel11lheory abom the g<><.\s which, they knOll, WeTC once inside the A'a. Acmrding: to theRurutan d,krs, there were thre.: g:o,h inside the .>\'a whcn it lias made, b\ a llero named .>\maiterni.'>\'l\"iter"i made the l\ 'a ,fter vi,iting, cit.\ nonc other than I-<mdon, present re'ting-pbce of the A'a,which he rea"hed in fulfilment of a specie. of knightly quest, impo,~J on him in order to II in the handof th~ adoptcd daughter of the King of Rumt", who had becn promi,ed (() hi, brother. In LondonAma;lel""Ji cnm"ntered thc God of Wi,dom (who bter \\""s thc G(}{I of the Chri,ti""s. brought toRun"u by the missionarics) whose imuge he replicatcd in the form <>f the f,mou. A·a. The god, insideth~ A'a were three Polyne,ian gods origin.ting in London: Ronm_enta_ore, ulias Te Ama I\lemo, ali"God the Father; Aura-roite"t", alia, Te Atu" lamaiti, ali" God the Son; amI Te at,la "i1Cm". "lias TeAn" Vam" I\"'ito'i, nli"s God the Hoi\" Spirit. tn other word<, the A', i, thl' Tabern"c1c in which theTrinil}' arriycd on Runn", by the agcn~,. of. R"rm;1n hero, tong bcture th,' mission,r;", thcm,eke.'~ni\-ed. The i\'a i~ in London. but;1 is pr",~nt on l{ur"llI in the form ofChriMian belief.

7.11. COl/centric idols and Frat/al Personhood

To exhibit the animation of the idol through the congruence hetween theexternal relational contcxt within which the idol is set, and the internal nexusof relations between rhe mind and the body (as a relation between inner andoutcr 'persons') consider the Polynesian example in Fig. ].II!I, a carving fromRurutu in the Allstral.!sles, which has been in London sincc 1822 and whichcan he seen at the Museum of ~,1ankind. This carving, in Rurutu called A'a,but more commonly identified as T<lngmoa, is arguably the finest extant pieceof Polynesian sculpture. Almost every other Rurutan idol was consigned to theflames by the missionaries, but this one was preserved, initially to drum upsubscriptions for rhe London Ivlissionary Society so that they could afford todestroy other, no doubt equally finc, carvings elsewhere.

Thc most striking attribute of this carving is the way in which the featuresof the god are represented by litlle figures which repeat, in miniature, thcoverall form of the god as a whole. This god sprouts little gods i11l over itssurf:lce: mathematically, it is <lkin to the type of figure known as a 'fractal', afigure which dcmonstrates the property of IelF-sillliliari~yat different scales ofmagnificationJminification. lvloreover, besides heing a god made of many gods,the A'a is also a box or an ark. It is hollow inside, having a lid at the back,and it originally contained twenty-four or more additional, smaller images ofRurutan gods, which werc removcd and destroyed in 1822. For all we know,the gods inside the A'a werc Themselves hollow, though I think not. Butwhereas we think ofhoxes as less significalllthan their contents, the A'a, eventhough it is a box, is the primary image of Rurutan Jivinity, encompassing andsubordinating all the subordinate gods who sprout from its surfacc and onceresided in it's interior. According to contemporary RurUTall traditions, theextcrior gods encompassed by the A'a correspond to the kinship units (clans)comprising Rurutan society as it whole." .Many other important Polynesian

Tit" DI.rlrih/lfed PersunIJ6

13H The Dislrib/lltd PnslIIl "I'll<' Distributed PerSOIl 139

Fla. ,. I1/1. The fracl~l goo: A'~

from Rurul~. SO/lru, The Orilishi\luscum, .\1.1\1 0"977

"IG. ;.11/2.Geneal"gicJ'personhood objt"£lified:a st'lfTgod frl)llllhcCo"k Isl:lnds. Sallrre:i\luscul1l of Archaeologyand Anrhrupulo;,:),Cunbridgc, Zoo)')

can iogs represent personhood in Ihe form of gene:llogy, as for instance therelated carving, a 'smIT goo', 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' di,·inity represents divinity as an assemblage of rela­tions between (literally) homunculi. In so doing, the A'a 00' iatcs the con­rraSl between one and many, and also between inner and outer. The surfaceof this image consists of amalb":lmated replications of itself, or ahernati"e1y, asuccession of budding prorube.....mces. Internally, the image consists of il<;c1f,replicated on a smaller scale, within irs own imerior cavity. As such, it imagesboth the notion of personhlXJd as the aggrcbtate of extcrnal re1:ltions (the Ollt­comc of gcnealogy, fanning Qui in time and space) and at the same time thenotion of personhood as Ihc possession of an interior pcrson, a homunculus,or, in this inst,lncc, an assembh\ge of homunculi. We cannot individuate theA'a in the way in which we normally individuate persons by identifyingthe boundaries of Iheir person with the spatial boundaries of their bodie.<;, forthe A'a has no such boundarics; it is like a Rus.<;i:m doll, and in this respect,it irresistibly recalls the lines in Pur Gynl in which the hero compares Ihe(moral, biographical) pentOn 10 an onion, composed of a succession of concen­tric layers:

-Why, you're simply an onion-and now, my good Peler, I'm going to pccl youand tears and entrc;llies \\'on', help in the least.rTahllg 11I1 (lIIio/l, lit sirips it skill by shnJThere goes the battered outer layer-thai's the shipwrecked man un the dillghy\ keel.This layer's thc p:lssenger-----scrawn) and thin,But still with a bit of a taste of Peer G) nt."ext underneath comes the gold-mining Sclf­lhe juice, if it c\'(:r hud any is gone.This rough skin here, with lhe hardened palt:his the fur-tr';1pping hUnler from Hudson's Bay.We'll throw that away without a word.~ext lhe archaCQlogisl, short but yigorousjand here's the prophet, juicy and frcsh-it stinks of lies. as the saying goes,and would bring tears HI an honest man's eyes.This .~kill, curled and effeminate,is the gtntlcm:m li\'inl; his life uf pleasure.The next louks IIllheallhy and streaked with blaek­black could 111(:,111 either priests or nil;gers .[He peels OjJset,'ffil/ layers at once.]Whal an incrcdible number of l..)crs!Don't we get 10 the hearl of it soon:

Prom the anthropologi(;al point of view, if not the philosophical one, thesolution 10 the confli(;t between the external notion of agency, deriving frominsertion in the social milieu, and the 'inlernalist' theory of agency, deriving

A fnctal person is nevcr a unit standing in relation 10 an aggregate, or an aggregateStandinl; in relation to a unit, but always an entity with relationship inlegr:lllr implied.Perhaps the most coner~te illustration of integral relationship eomes from lhe gener­alised notion of reproduction :mc.l genealogy. People exist reprodueti\'c1y by being'carried' as part of another, and 'carry' or engender others by making" themseh'csgenealogica.l or reproductive 'factors' of these others. A genealogy is thUS:lll enchain­ment of people, as indeed persons would be seen to 'bud' OUI of Dill;: another in .1~peeded-up cinematic depiction of human life. Person as human being and person aslineage or clan He equ:111y arbitrary sectiollings or identifications of this enchainmcnt,different projections of its franality. But then enchainmCIll through ?odil~' rcprr~du~­

tion is itself mcrely Olle of a number of instantiations of integral rciarumshlJl, whIch ISalso manifest, for i/15t,l1lCe, in the commonality of sharcd language. (199 r: 16])

PL'Cr GYIll'S onion is also a fractal, of the same essemially concentric form asthe A'a. Ihsen's idea, in utilizing this image is to show that there is no ultimatebasis to Peer Gym's personhood; he is made of layers of biographical (rela­tional) expericnce accreted together, for which none the less, he must ('ake soleresponsibility. I'erhaps it is not such a vast step to pass from Peer Gym trappedin the aporias of ninetecnth-ecntury materialism and individualism to the theo­logical impulse whil.:h motivates the A'a, which depicts the divine CrL'3tor, themind of which the world is the hody, in the form ofa body composed of otherbodies, ad inji/lilllfll.

This idea is given contemporary expression in the work on personhood inMelanesia, by writcrs such as Marilyn Strathern (1988; ef. Gelll9(8) and RoyWagner (1994). Wagner, in particular, has developed the notion of 'FractalPersonhood', which he mobilizes to overcome the typically 'Western' opposi­tions between indi\'idual (ego) and society, parts and wholes, singular and pluraLThe notion of genealogy, which is so signally expressed in our twO Pulynesianexamples (both idols, of coursc;), is the key trope for making plurality singularand singularity plural. Any individual person is 'multiple' in the sense of beingthe precipir:ue of a multitude of gencalogical relationships, e:lch of which isinSTantiated in his/her person; and conversely, an aggregate of persons, such asa line:lge or tribe, is 'one person' in consequence of being one gcnL-alogy: theoriginal ancestor is now instantiated, not as one body bur as the many bodiesinto which his one body has transformed iTself. Wagner writes:

141

from an inncr subjeclive self, is to be sought in this 'cnchainment', the srruc­tural congruence between the inner self (whi(;h is relational) and the outerself (which is equally rc1ation:l1, but on an expanded scale). The 'genealogicaltheory of mind' which is explored, particularly, in Strathern's work (1988)seems perfecdy expl'cssed in the form of the A'a.

But this artwork, and its Polynesian cousins, is nOI unique. In f.,ct, thepossession of 'signifieJ.nt interiors' is a \'ery common fearnre of sculpturalworks specifically intended for cult use, rather than as mere representations.Sculprnral images which open up, like the A'a, 10 reveal other images, weremanufactured in ancient GreeL'C, and provided Akibiades with a simile for hismentor, Socrates (as narrated by Plato in the SympoSiU11l2IJ-15):

I am here to speak in praise of SocraICS, Gcm1cmen, and I \\ ill just do it by means ofsimiles. Oh yes, he will perhaps think it is only for a bit of fun, but my simile willbe for truth, not for fun. I S3y then, that he is exactly like a Silenos, the linle figures[hat you see in statuarics' shops; the craflSmc:n make them, the~' hold panpipes or pipes,and they can be opened up down the middle or folded hack, and then thq' show insidethem, images of the gods. And I S3.y further, that he is lile .\ Iarsras the Satyr [who wasflayed by ApolloJ ...

Here the OOntra."t is between the ugly exterior (body) and the divine interior(mind) of Socratcs, The same kind of imagc was also developed in Christiancult art, though with a different theological implication. A class of holy statuescalled 'vierges ouvrantes' was made in the l\liddle Ages, though few havc sur­vived into the present, perhaps because these images wcre particularly con­ducive to idolatry, as the following p3ssage from Camillc's work of The Cothi,Idol shows. 'Our Lady of Bolton' which llsed, beforc the Reformation, to standin a chapel of Durham L-athedral was:

a man·eylous Iyvcly and bc\\lifull Image of the picmre of our Ladie, so calld Lady ofRoultone, II hich picture II":IS made co open with gymmers lor two le:.vesJ from herbrcaSL<; clowlI\\ards. And within this saic.le immage was wroughte and pictured fheimm:lge of our Saviour, merveylouse fYlllie gilted hold)'llge uppc his hands, and hold­ing betwixt his hands a fair large Crucifix of Christ, all of gold, the which Crucifix wasto be taken fourrhe e\ery Good Frid:lie, .llld C\'ery man did creepe unto it that was inthat Church at that daye. And every pril1cipall daie the said image was opened, thate\'cry man might sec pictured within her the Father, Son and the Holy Ghost, mostcuriollslye and fyndy gilted. (C:mlille 19SIj: 230-1)

Camille illustrates a German example of such :t 'vicrge ouvrante' (Fig. 7. t 1/3).'Our Lady of Dolton' was related 10 the morc COlllmon type of Christian holyimage inlhe form ofa reliquary, in that she waS, besides an image oCthe Virgin,a corl1ainer for the goldcn crucifix which Was paraded on Good Friday, (Foran Indian parallel to a 'vicrgc ouvrante' d". Fig. 7.11/4 showing I Ianuman,the monkey god, opening his bre:lsl 10 re\'eal Rama and Sita.) Such imageswere contrO\'crsial even whcn they wcre still in commun lise: Camille quoles

(Ibsen H)66: 191)

Th,' Distributed Person

rHe pulls {he wlw{1' Imirm t/J pie((s.l1\0, I'm damned if we do. Right down to the ccntrethere's nothing hut layers-----smal1er and smaller.N:lture is witty!

140

142 TIle Distributed Penon 'lI1,; Disrrihll/ed Penol/ 143

rIG.7.lIh, 'Vicrgcouvrantc'. Source: CamilleHJSl): pI. [2+ NurembergGcrmanischcs National]\Iuscuill. Painted wood.H, 126 em

the theologian Gerson, in 1402, denouncing 'Carmelite' images which 'ha\'cthe Trinity in their abdomen as if the entire Trinity assumed flesh in the VirginMary. .' (H)81): 2]2). These images were all too animate, they smacked ofnecromancy rather than religion. Later on in the same book, Camille showshow reliquary heads, containing the bones of saints, could become objects ofJeep official suspicion, especially the silver and golden heads supposedly wor­shipped in pag".ln ceremonies by the Knights Tcmplars, who were suppressedin 1308 as a result (ibid. 271---'7). The notorious 'talking head' reputedly devisedby Roger RIcan is another varilltion on the same theme (ibid. 246-7). But therelation between religious art and sorcery runs very deep, we have already seen.

No such opprobriulll attached to more ordinary reliquaries in human form, ofwhich there are mailY famous examples, such as those illustrated and discussedby Freedberg (1989: (}2-S, figs. 30-2). This author also remarks on the factthat in the j\'Iiddle Ages, churches could not be consecrated at all unless theyhad holy relics installed in them. Just as relics animated the reliquary image,rendering it a particularly holy object (an object wirh a mind, or perhaps moreprecisely, a spirit, within it) so lhe church (fabric) as a whole became a 'body'whose animation also required the insertion ofa relic. The normal place for the

FlG. 7.[1/~. An Indian parallelto 'vierge ouvrante': themonkey god H,lnumun rtvtalsRamu and Situ in hi, brtasl.Source: i\, Mooktrjtt ([ 980),Ritual Art of I"dia (London:Thame, & Hu<hon), pla(C 80.Kalighat school, southCalcutta, L 1850. Gouacheon paper

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

The insertion of animating relics into images raises a fresh question, how­ever, 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 oIConsecratiol/

his f.1ir to say that the worship of images or idols is most eXlensivcly prac­tised, nowadays, in south Asia, among Hindus and (in a rather marc qualified

144 The Distributed faSQn 14S

way) among Buddhists. Here images are still being produced and installed inholy places in great numbers <Jnd here it is possible for amhropologists to

observe the rites of consecration in detail. I shall therefore conduct the argu­ment of this section with reference to three well-mo\\ n south Asian examples,the consecration of the idol of Jagg-.math in Puri (F_-.chmann d al. 1978), theaccount gi\'en by Richard Gomhrich (1966) of the.inStallation of a statue of theBuddha in a Sinhalese monaster)", and finally, the aceoUnl gi\'en by ~lichael

Allen (1976) of the conseC•.lt'ion of the '!i\'ing image' of the goddess Ta1cju(Durga) in Kathmandu.

The images of Jagganath, his brothers, Balabadhra and Subadhra, and hiswife Sudarsana in the temple of Puri (Orissa) are among the most re\-eredidols in the whole of India, but, though anthropomorphic, these image... arc(\'isually speaking) quite obviously cylindrical sections of tree-trunk, drcssedup and equipped with vestigial upper limbs and very, vcry, large eyes. Theimages arc not old in physic.11 terms, though their design is indeed ancient,since rhey have to be renewed every twelve, or at mOSl every nineteen years,in the course of a ceremony called I/tiVllkalevllra. A derailed description ofthis ceremony (contributed by G. C. Tripathi) is provided in Eschmann'swork on the history and affinities in 'tribal' teligion of Lhe Jluri Jagg.lnath cult.The lribal affinities o( this imPQrtant Hindu cult arc not in doubt; I myselfworked in the same region of India (Bastar district. which abuts onto the OriSSo.1highlands to the west) among: tribal Muria Gonds, whose images of diyin­ities were aniconic wooden POStS, without limbs or eyc..., but otherwise highlyresembling the Jagganath imagl.'S, though on a rcduced S(.~le (A. Gel1 1978; cf.S. Gell 19(P).

Tripathi's account of /J(lt'aJ:alrram is a first-rate exercise in ethnographicdescription, which I cannot unfortunately summarize in any detail. The cere­mon ~ of renewal has fiye phases, as follows:

I. To find out the dorn (sacred wood) with the pr~rihcd characteristicsand to bring it to the temple (ill\-oh-ing a sacrifice to ward of e"il spiritsetc. and to sanctify the trce before felling it);

2. Thc l~n-ingof the wouden structure of the images;3. The consecration of the images by lhe insertion of the 'life-substance'

(lmtlwwp{ft!arl!t{f) into them;4, The burial of the old figures, the funeral and purificatory rites of rhc

Daitas (temple servitors, of low caste);;,' Giving the images their final form by means of several coverings of cloth

etc. and by applying p:lil1t to them (Eschmann 1978: 230).

[ omit the first two phases in this process, except 10 mention rhar the efficacyof the images depends, initially, on the auspicious location and form of the damtree (it must grow by \\'.uer, be surrounded by three mountains, have dark,'rcd' bark, a straigln uunk, with four branches, etc.). J'Ull1crous ceremonics

accompany lhe felling, transport, and carpentering of rhe ,Iaru trcc. The imagesarc made by temple servitors of low caste but high ritual privilege, the Daitas,ex-tribals who alone know how to create this type of wooden image, accordingto precise rules, import:mt aspects of their \\ ork being kept secret by them,even from the Brahmin priests. Among the most important of their secrets isthe actual nature of rhe 'Iife-substance' (bmlmlllpIIJarlha) of the images.

The consc..'Cration procedure Ulmmenct.'S simultaneously wirh the carving ofthe images, and is first of all conducted by 13rahmins. Since the images arc stillbeing made, the Brahmins demte their cfforts to consttrating a separatc pieceof daru \\ood, which, divided into four pieces, will become the 'lids' O\'cr theC3\'ities in the finished images containing their 'Iifc-suhstance'. This piel.'C ofwood is protected with offerings to e\·il spirils, Ihen elaborately bathed andpurified, then placed to rest on a ritual bed. After this, the spirit of the godr\arasimha (of which Jagganath is a form, as well as being a form of Krishna)is induced into the wood by the recit.lrion of mantras on each of its parts(equated with the parts of a body) over a number of days. This invocationprocedure is called 1I)'lIsadaru. After this. the ritually treated dam wood andby metonymy, the larger sections of dam wood which are at this momentreaching completion as images in the Dailas workshop--is spiritually speakingendowed with life, flesh, hlood, sense organs, etc. It can then be cut into fourlid-pieces to fir into the (our images.

The crucial ceremony however, is not conducted by the Brahmins but bythe Daitas. These take the 'old' images frOIll the temple, and strip from themthe many layers of resin-impregnated cloth with which they are hound. Theycan thcn reach the compartments inside Ihe old images in which their 'Iife­substance' is secreted.

The Daita entrusted with the job opens the belly of the old image in dead of night \\ithhis eyes blindfolded and the hands wr:lppcd up to the elbows so that he lllayncil1t~r see nor feel the brllhmllpadiirtha of the image. llte casket containing thcBrahmapadilrlha is then ul.:en out of the old muni [imageJ and placed in the ne\\ onc.The em'it)" of Ihe nCI\ image is then COl ered II ilh one of the four pieces of the;:-.J~asdaru[sacred \u.lodl which has been consecrated for aoom two wccks by the Brahmins,(Tripalhi in ESt:hmann un8: 260)

J'obody really knows, except the Daitas, exactly of what object or substancerhe brallllwpadarrlw is composed. It may be .t relic, a portion of a lloodenJaggamnh supposedly incinerated (but not completely) by the Muslim icono­clasts under Kala Pahada in 1;68. In the opinion of the Brahmins, it is asillagrall/a, .1 type of sacred ston~, uSlIally all oval river-pebble from the l-lim­alayas cOlllaining f(lssilized ammonites, and/or cavities (sec Fig. 7,12/1 fromMookerjee and Khanna 19/i).

The Daitas now pert(mn burial ritcs over Ihe deceased im<lges which havelost their life-substance. They weep and mourn, observing mortuary PQl1utionrestrictions for ten days. I Iowc\"er, the~ deri\e benefit from this, in that lhe}

146 The Disllibu/<'Il Person The Distributed Person 147

FiG. j,12/1. 'World-~~~d':

Jiliagram". Sounr: l\'1ookcrjccand Khanna '977

are allowed to keep the cloth wrappings of the old images, which thev cut into~trips and sell to pilgrims. These pieces of doth confer protection ~nd goodtortune on the purchasers.>

The final phase of consecratiun now occurs. Another caste of temple ser­vants, said to be Kayasthas (a low form of Brahmins) undertake the wrappingof the images-who, in wooden form arc just 'skeletons'-with their 'flesh',thaL is, with cloth and paint. The wooden 'bones' arc first of all 'washed' bvbeing impregnated with camphor oil. This gives the hones 'marrow'. The~,apparently, long red threads arc wound around them, representing 'bloodvessels', After this, many strips of red cloth (for Oesh) impregnated with resin(blood) :ll1d starch (fat, semen) are added, till the image begins to assume itsfinal form. The ollter wrappings are the image's skin.

Finally, the images are painted by craftsmen skilled in this art (rhital<ara).The very last act, which finalizes the consecration of rhe new images, is thepainting in of the pupils of the immense eyes of the images, which is done bythe Brahmin priests themselves, reciting Vedic mantras the while. After givingthe last sTroke of paint to the eyes of the images, the Brahmins give each a bath,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 kept for

, Th;s [',e,ogat;,.. ,ecall, the 'fe.thcr exchange' following Ihe decortication of lhe Tahir;an/O '0 (7.6,oo"e), 'l'he ",rmu ofJoggollalh i. disseminalcd ,·i. his eX""i"e, his body-part"jwha (le"'ings) erc.Incidl".lllall}', rhe SamC is true ill Ihe Egypl;an example di,cussed in /.10 .bove. E'ery time the id"ls'dothe, were renewed, rhe di«:"nlcd IrMments were di.trihllled 10 imporTant person., 10 use 10 wrapIheir 0"''' corpses in", mllmmy-c1olhs. Where'S in Polynesia "nd India_ the cloth/feather exuviae of thegod well! to henefit the living. in Eg) I'r they henefited the dead in the "fterlife,

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 :L';pects of these interesting ceremonies arc of particular COncern to us,The consecration of the images evidently proceeds according to two parallelstrategies simultaneously; first of all, the strategy of the Daitas, which focusesparticularly on the placing of the life-substance inside the image, in the cavity,and secondly, the strategy of thc Brahmins, which proceeds in the inversedirection, through the apostrophization of the representative billet of wood withlife-endowing mantras (the phase of nyasadllru) and the application orthe finalstroke of paint to the pupils of the images' eyes. In other words, the Daitas' pro­cedure could be called the inside---out procedure, while the Brahmins' procedureis the oUlside--in procedure. Both are necessary and mmwllly complementary.

The Daitas' procedure, is, perhaps, the more primitive, in that it seeks ~orender a physical analogy between the 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.The Brahmin strategy, for the most part, is more abstract than this; ratherthan fashion the image, they address it with mantras, animating it externallythrough 'the magical power of words' (Tambiah H)8S). We can easily see rhatthese two strategies correspond exactly to the two strands previously identifiedin the philosophy of mind ,lOd agency. The Daitas' strategy is the 'internalist'one, the Brahmins' the 'externalist' one. But it would be false to supposethat these two strategies arc independent. In the end, the Brahmins concede tothe internalist strategy to the extent that their final act is onc of physical modi­fication of the image, not mere aposrrophization, Their culminating act ofpainting 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 holes, orifices, giving access to the hidden interiorwithin which 'mind' resides, The surface of an idol is not an impermeable bar­rier, 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 unliT

rtln never be reached. The outermost (relevant) skin of rhe idols consists of thetemple of Puri itself (which is, of course, a microcosmos) which is filled withsacred words and odollrs-verbal and olfactory skins (cf. Anzieu 198y: 59).The idols reside in the centre of, and animate, this reverberating microcosmos,and are animated by the incessant Oow 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 devo­tions, animate the idols occupying the cynosure ~r a great assembly of souls.Thc idols themselves arc enshrined at the centre, framed on their altars,adorned with masses of flowers and jewellery, presiding over their material

148 The Distribuled PerSOi/ The D!~'trillllfi'{1 Persull 149

wcahh of hcapcd~ur otTcrings, their external skins in rhe form of possessio/Js.vVe cannot approach or touch the idols, so we can proceed further on our jour­ney towards the centre in imagination only. What we see are their visible skins,bur these are only Olller wrappings. These wrappings consisl of numerouslayers, their inner skins of flesh, fat, semen, blood, bOlle, marrow. These weIllay penetrate, one by one, or gazing into the enormous eyes of the idols, wecan enter their bodies directly. But whal is there, concealed beneath, behind,the inky pools? Inside, [here 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 inrhe casket? Even the man who placed it there does nor know perhaps; he hascertainly never seen its COnlents. 'Ve believe that there is:l mlagrtlfll(f there, asacred stone or world-seed. [f so, whar is in the salagrtlflla? ·fhe salaJ;mlnll itselfhas an imerior, and, holes leading into this interior (cf. Fig. 7.12/1). We mustenter these holes. And thcn what would we find? Who can say-and does itmaner?-for by now it is apparent that the animation of the image is not a mat­ter of finding the 'sacred centre' at all. "rhat matters is only the reduplicationof skins, outwards towards the macrocosm and inwards towards the micro­cosm, and the filet that all these skins are structurally homologous; there isno definitive 'surface', therc is no definitive 'inside', bur only a ceaseless pas­sage in and out, and that it is here, in this traffic to and fro, that the mystery ofanimation is solved.

The consecration of an image of the Buddha in Sri Lanka, dcscribed hyGom brich (11)66; cf. the discussion in Freedberg 11)81): 84-'7, (5) follows a sim­ilar pattern, though reduced. Buddhists, especially monks, are not supposedto worship idols, but showing respect to images of the Buddha, by makingofferings and gestures of submission, is one way to acquire merit and secure agood rebirth, if not a vcry significant one. Once again, it proves to be the casethat images of the Buddha are, if they arc to havc any religious importance, alsoreliqwlries..Minute portions of the Buddha's bodily remains are placed insideimagcs to render them eAlcacious. However, this docs not consecrate them,according to Gombrich; consecration is accomplished by the craftsman, whopaints in the eyes of the image in the course of a special ceremony caned netra

pinkama the 'eye ceremony' (H)66: 25)·The ceremony presents some interesting contn1sts to the consecration of the

Jagganath images, as well as many points of continuity. The BrahminfDaitarelationship is partiany inverted in the ritual division oflabour in the SinhaleseBuddhist consecration ceremonies. In Puri, it is the Daita who anim:ltes theimage by putting 'life-substance' into it, and the Brahmin priest who paints inthe eyes; in Sri Lanka, on the other hand, it is the monk (abbot) who pbces therelic into the Buddha statue, and the lay craftsman who paints in its eyes. Thisis a precise reflection of the difference between the Buddha (a dead humanbeing with morally supernatural characteristics) and the Hindu gods, who are

non-human immortals. A Buddha statue celebrates the possibility of 'a gooddeath' and monks are semi-dead individuals W110 aspire tu the uhimate gnod­death condition. Consequently it is only appropriate that the handling ofBuddha relics, which arc pieces of a dead body, should be assigned to monks,who are semi-dead themselves, and who of course always preside over fun­erals (but nor births and marriages) in Buddhist cOllntries like Sri Lanka (orThailand; cf. Tambiah 1(85). In a sense, then, what the relic docs 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 can see that this ritual procedure hardly amountsto the 'animation' of the image. Gombrich understands this as the means of'legitimating' the statlle as a Buddha image, so that from the monkish per­spective no taint of idolatry arises. For the laity things are otherwise. Buddhais really a god, and approaching the srame of the Buddha, the lay worshipper,immersed in lite and sin, seeks personal reassurance through communion withthe Buddha's living presence, conveyed thrnugh the eyes (see above on theHindu equivalent of this form of worship, darshan). Because only lay peopleare 'superstitious' enough tu engage in such theologically decried acts ofwor­ship, it is the lay craftsman who is charged with t'he task of animating theBuddha image.

The ceremony is regarded by its performers as vtry dangerous and is surrounded wilhtabus. It is performed by the craftsman who made the ~tatue, after several h()ur~ ofeeremonie~ to ensure that no evil will come to him. This evil, which is the ohject ofall Sinhalese healing riruals, is imprecisely conceptualised, but results from makingmistakes in ritual, violating tabus, or otherwise arousing rhe malevolent atttntion ofa supernatural being, who uSlwlly conveys the evil by a :;aze (bii/mli). The craftsmanpaints in the eyes at an auspicious moment and is left alone in the dosed temple withonly his colleagues, while everyone dsc stands clear even of the outer door. Moreover,the craftsman does not dare to look the statue in the face, bur keeps his back to it andpaints sideways or over his shoulder while looking into a mirror, which catches the gazeof the image he is bringing to life. i\S soon as the painting is done the craftsman him­self has a dangerous gaze. He is led out blindfolded and lhe covering is only removedfro111 hi~ eyes when they will first tall upon something which he then symbolicallydestroys with a sword stroke. (Gomhrich 1966: 24-5) lThis could be an animal, suchas a bull, bUl a pOL, or a lree which exudes sap can be substituted.l

The detail of the mirror recalls the use of bronze mirrors by rhe Brahminptiests of Puri to 'bathe' the images by splashing water over their reflections.But the Puri Brahmins are unafraid to look at the images directly while paint­ing in their eyes. Unlike thc eyes of Jagganath, the eyes of rhe Buddha, whenfirst opened, innict death on the very person who performs the opening (a caseof Anist-A ------? Artist-Il; sec above, Sect. J.I J). The craftsman must there­upon slay somc creature in order not to die himself. This is vcry f;1I· fromtheologically Buddhist, and in olden times, the craftsman was drcssed as a

150 The Dis/I'ibulal Person The Dismbuled Person lSI

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

freedberg (1989: 95), commenting on Richard Gombrich in his magnificenttreatise on the reception of religious and other types of images, raises a verybasic question. Are imagcs such as these powerful and efficacious (religiously)because of the ceremonies of consecration which have endowed them withsignificant, occult, characteristics, or is it bec,luse 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 been shaped and formed by art into the semblance ofpersons, deities, etc. Such a point of view is necessary, ,md no doubt proper,for the art historian, who has to distinguish the 'power of images' from thepower of mere unformed things, however sacred and sacrifying their origins.The anthropologist is in a slightly different position, however. The 'objectswhich resemble human beings' with which the anthropologist deals, prim­arily, are not portraits, effigies, idols, and so forth, but simply human beingsthemselves. Freedberg's emphasis on the centrality of artistic mimesis has,anthropologically, to be sct in the context in which the 'represcntation' of ahuman being, or indced 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 imagc inperforming the mass and uttering Christ's words. This is not to dismissFreedberg's questiun, to which I will return in the next secrion, but beforele'lving the subject of consecration, ir is interesting to compare the two con­secration sequences wc have examined, with a third, in which the 'index' ofthe divinity is not a carved image at ,111, 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 particularfeature of the religious system of the .Kewars of the Kathmandu valley, wherethe cult has been the subject ofa detailed study by IVlichael Allen {1970}' Thereare some nine or ten living goddesses in the l\ewar region, of whom the mostimportant is the one traditionally associated with the royal household. The \'irgingirl is a form of the fierce royal deity, Taleju, who herself is a form of Durga,the violent and erotic goddess in the Hindu pantheon, the slayer of rhe bufhlo­demon who rides on a lion and brandishes a sword. Allen's analysis concentrateson the paradox whereby a virgin, premenstrual girl comes to represent so fear­some a divinity, but what concerns us is only the mechanics of her divinization.

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

the deity, in this form, is temporary and generally ecstatic; the medium goes intotrance and becomes a 'horse' for the deity, making utrerances 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 Rumariis and remains the goddess in person until the moment of deconsecrationwhich transpires when certain 'negative signs' arrive-loss of milk teeth intheory, in pracrice, menstruation. The kumari comes from a particular caste,who arc Buddhists, traditionally attached to the monasteries of Kathmanduwhich were disbanded by the present royal dynasty. She is, in caste terms, non­polluting hut somewhat outside the l-lindu hierarchical systcm. 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, especiallywirh regard to the king. I-Iere is Allen's account of the circumstances of herinstallation, which occur at the end of Dasain (Dassara) the festival of Durga:

At nightfall eight buffalocs representing the demon an: killed by ha\'ing 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 buffaloes and 54 goats are killed in a similarmanner. As may well be imagined, the small (ourtyard [of the Talcju templeJis by thenawash with blood ... At this point, usually about 1.00 A.M. the small Kumari-e1ect isbrought to the entrance. She is supposed to walk by herself: in a elockwise directionaround the raised edge unril she reaches rhe bloody Taleju shrine. She must enter it,still maintaining a perfecrly calm demeanour, and if all is well she is then taken upstairsto a small roum fur the installation cercmony ... after the uSllal purificatory and otherpreliminary rites, the (hief priest performs the main eeremonv in which he removesfrom the girl's body all of her previous life's experience w that the spirit ofTalcju Illayentcr a perfenly pure being. The girl sits naked in front of the priest while he purifieseach of her sensirivc body areas in turn by reciting a mantra and by touching e,lCII areawith a small bundle of such pure things as grass, tree hark and leaves. The six sensitiveparrs are her eyes, throat, breasts, navel, vagina and vulva. As he removes the impur­ities rhe girl is said w steadily become redder ,llld redder ,IS the spirit of the goddessenters into her.

At this stage the girl is dressed and made up with Kumari hairstyle, red li1>a, thirdeye, jewellery, etc., and then sits on her beautifully carved wooden thrune un the scatof which The priest has painted the power/tIl sri .J!anfra mal/dala of Taleju. She alsoholds the sword uf Taleju and it is at this point that the final and (omplete transfor­mation takes place. It is worth noting rhat though from now until her disqualificationsume years laTer she will be continuously regarded as Kumari, it is also believed that it;s only when fully made up and sitting on her throne that identifi(ation is (oTllplete. Atorher times, especially when casually playing wirh friends, she is parrly herself andpartly Kumari. (Allen lIJi6: 306-7)

To what extent can one detect a parallelism between the installation of thekllmari as a 'living icon' ofTalcju, and Ihe inSl'aliation rites of more convent ional

152 The D1Ilribuied Person The Dis/nhulal Per,.oll 153

FIG, 7.1ZI2, KUlllllriwith paintedthird eye. S~uru: Allen 1976

idols? As with each of our previolls examples, the consecration of the klllllariproceeds in two phases, one focusing on her interior and the other focusingon her exterior. The first phase consist of the 'emptying' of the kumari-electof her past life (i,e. her personhood, agency, as a mere human heing), whichis objectified as the impurities removed from her orifices, eyes, tlHoat, vulva,etc. She becomes a 'hollow vessel' into which, through the extraction of allprevious contents, new contents may be drawn; that is, the spirit of the god­dess Talejll. This phase, it seems to me, is the equivalent of the phase in theconsecration of the Jagganath images in \vhich the Daitas hollow out a cavityin the image and place therein a (foreign) life-substance. But after this has beeneffected, the transformation, as Allen makes clear, is still not complete.

'fhe second phase of consecration consists of 'wrapping' the t..'umari in thedress of Taleju, and painting her, as well as providing her with the goddess'sprimary attribute, 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 oCher own. But they outline her existing eyes with exag­gerated make-up, besides adding an enormous painted eye above them (secFig. 7.12/2). The parallel between the making-up of the kumari with an extraeye and the painting-in of the eyes of con ventiom! idols is surely rather srrik­ing. 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 inrhe usual way by the recitation of sacred words, and the metonymic effect ofbeing scated on a magical design of great power, the sri )'alltm manda/a.

In short, there is little to difl"crenriate the consecrntioll of the kumari fromrhe consecration of any other idol, except that the kumari can walk, and talk,

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 workthere is an insensible transition berween '\~'orks of art' in artefact form am!human beings: in terms of the positions they may occupy in the networks ofhuman social agency, they may be regarded as almost entirely equivalent.

7.J3. COl/rlllsiof/: Fromlhe Indh.Jiduul to the Collective

Thus I conclude this extended discussion of idolatry. I recogni'l.c, however,that the particular line 1 have taken has consistently resulted in the emphasisbeing placed on sociological, religious, and psychological agency at the expenscof aesthetic and artistic agency. And r am left with Freedberg's pertinent ohjec­tion: to travel too far down this road is to lose sight of art's specificity. While Ihold that where each individual work of art (index) is concerned, anthropolog­ical analysis is always going to emphasize the relational context at the expenseof artistic or aesthetic form, the net work of agent/patient relations 'in the vicin­ity' of the work of art-the same docs not apply when we corne to consider art­works, not 'individually' but as collectivities ofartworks. So far, each index that1 have subjected to analysis has been regarded as a singular entity, embeddedin a particular SOCi,ll context. However, artworks arc never just singular enti­ties; they are members of categories of artworks, and their significance is cru­cially affected hy the relations which exist between them, as individuals, andother members of the same category of artworks, and the relationships thatexist between this category and orher categories of artworks within a stylisticwhole--..a culcurally or historically specific art-production system.

Artworks, in other words, come in families, lineages, tribes, whole popula­tions, just like people. They have relations with one another as well as with thepeople who Cl·eate and circulate them as individual objects. They marry, soto speak, and beget offspring which hear the stamp of their antecedents. Art­works are manifestations of 'culture' as a collective phenomenon, they are, likepeople, enculturated beings. So far, none of the collective issues surroundingthe work of art have been considered. In order to broach these issues, it isnecessary to adopt a new register. Here, I can make ,lmends for having writtenso many pages which may have seemed tangenti'll to the study of works ofan as normally understood. I shall, for II while, desist from the terminology of'indexes' and abductions of agency, and suchlike, rcvcrting to a more conven­tional vocabulary. Because there is one 'conventional' art-theorerical conceptwhich even the most radical anthropologist of art cannot pur to one side-theconcept of style. Style, which is the harmonic principle which unites works ofart into groups, into collectivities, corresponds to the anthropological theme of'culture'. Culture is style, really, just as Fernandez suggested in a deservedlyinfluential discussion (1973).

154 Tht Distrihuud PtfS/m

Each individual work of art is the projection of certain stylistic principleswhich form larger unities, just as each indi\'idual, in a kin-based society, isregarded as a projection, into the here and now, of principles of desecm andalliance and cxchange. The concept of style allows us, for the fin;t time, toconcentrate exclusively on works of art as such, and to discuss whal mayappear, indeed, to be their 'aesthctic' properties. So let me reassure those ofmyreaders who arc aesthetes and art-lovers (if I have any left afler all the abuseI havc showered on their heads}-you have reached the one chapter in thisbook you might possibly enjoy reading.

8

Style and Culture

8.1. Oil/he Concept oISZlde

In this chapter, my aim is to formulate a COnl,:cpt of 'style' lldapred to therel/ui rernellts of the anthropology of visual art. In thc anthropology-of-<lrt con­text the concept of style is distinguishable ftom the concepts of style applic­able in Wcsrern art history .mel aesthetics in thllt the 'units' of style arc not(usually) individual artists, or schools of artists, or movements, but 'cuhurcs' or'societies'. j\crually, the units of style arc conventional ethnographic isolates asrepresented for study porposes hy museum collections and published sourceson material culture. Such ethnographic isolates arc historically bounded as wellas geographically bounded; usually the mllections and documents belong toa particul3r period. Where studies of ethnographic an in museums arc con­cerned, the period in question c~incides mostly with the colonial era. Thefocus is primarily on 'trnditional' art forols, though it hardly needs to be saidthat during the colonial period there were startling historical devclopments inthe so-called 'traditional' societics 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 discuss at thisstage; besides which the subject has already received detailed attention fromnumerous other researchers, most notably by 1'\. 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, sClling to oncside the acknowledged problems which these assumptions admittedl} f'"Jise.

The problems of historical contextualization do not materially afTect what Ihave to say about style, since the description of a style may be as broad or asnarrow in scope as necessary 10 accommodate any gi\en historical perspeCTive.The question that I want to address is not historical bm conceptual: what docsthe concept of 'st} Ie' contribute to the undersranding of material culture?'Style' is a vague \\'ord of uncertain definition and many. rather disparatc, uses.Finding a usc for it in the context of the anthropology of material culture mightbe considered a waste of effort, were it not, in f:1CI, so pervasive, at lcast as amode of classification. As it is, we are routinely accustomed to c1assif)'ingobjects as sharing, or not sharing, stylistic attributes with one another. Butexactly what is shared (or not shared) in such instances is much harder to ass~.

J\loreO\er, we are inclined to belien: that what objects with shared sl}listica\tributes ha\'c in common, is nOI just some formal, external, propert}, but

156 $Iyle lind Culture Style 1/1/(1 Clilfure 157

something integral TO their standing as expressions of'the culture' in the widersense; common stylistic anriblltes shared by artefacts are associated, via a basicscheme transfer, with shared 'cultural values' in a community.

[ believe that the intuition that there is a linkage betwl:en the con(:ept ofstyle (as a (:onfigumtion of stylistic auributes) and the concept of culture (as aconfiguration of intersubjcctive understandings) is well founded. But I'Cl con­vert this intuition into a series of explicit and defensible arguments is a diStantgoal, whi(:h can only be appro.'lched by inl.Temental steps. One of the first ofthese steps is to clarify the notion of 'style' and [0 dislinguish, from amongthe many possible interpretations that C'J.n be plal.'Cd on this concept, the onewe specifiC'J.lIy need in the anthropological context. No anthropologist, to myknowledge, has anempted a critical analysis of the concept of'style', and as yet,there is no distinctively anthropological stylc-concept as such. In this situation,the only possible procedure is to turn [0 the aesthetic and art-historicallitern­rure on 'style' so as to canvass the \'ariolL" possibilities.

A concept of style that we probably cannot make anthropological use of inirs original form is the one advanced by Wollhcim in his desen'edl~ well­known discussion of style in pictorial art (1987). I mean no disrL'SJ)(.'Ct toWol1heim in saying so; it is merely my conviction that aesthetics and amhro­pology arc very different enterprises. Even so, Wollheim's discussion is \'eryuseful because in excluding certain questions from the purview of stylisticanalysis-because they do not coincide with his own interest in art he suc­cinctly adumbrates perspectives collateral to his own. He distinguishes, firstof all, between 'general' (colh:tivc) and 'indi\-idual' style. He is interested in'individual' slyle (as he suggests all aestheticians ought to be). On the Olherhand, an historians (and anthropologists) have more reason to take general orI.:olh.:cti\"e style as their theme. Within the category 'general' style he furtherdistinguishes 'unl\'ersal' style categories, such as the one which opposes 'rep­resentational' (realistic) Style to 'geometric' style (non-realistic, abstract, etc.).Opposed lO 'universal' sryle categories are the general or collecrive styles ofperiods, schools, and so on-these he calls 'historical' styles. The contrasroctween 'universal' styles and 'historical' (or cultural) styles is relevant 10 theanthropology of art. Certain anthropological/stylistic categories are 'universal'in this sense, Ihat is, not confined to particular cultures or traditions, but man­ifesting lhemselves in disparate cultural contexts: ami an instance of such :l'universal' stvlistie trait is 'split represenration' (Bo:ls 1927; Levi-Strauss1963), which 'will be discussed at a bter stage. Other anthropologic,ll/stylistiecategories arc historil.:ally specific, or culturally specific.

Wol1hcim contrasts his own view of 'individunl' (painting) style, whichhe calls 'generative', to the view of style taken by art history which he calls'taxonomic'. According to him, nothing is explained merely by constructingcompcndin of the stylistic nuribmes manifested in the work of a panil.:ular

artist; this is lTlere classification or pigeon-holing. To be explanatory, a stylisticanalysis must bring out those aspects of a painter's work whieh arc especiallypsychologically Julient. 'Style' is equ;ll'cd with psychological saliency, the capa­city, possessed only by p,linters with a de\'c1oped personal style, to so engagethe spectator's attention that the aesthetically significant as[X:cts of the workof art are the ones which actu:llly do attract our notice. Picasso's style, so tospeak, is Picasso's ability to ClUse us to notice his Picasso-ish nrtistic intentions.Style is personhood in acsthetic form, a 'precondition for aesthetic interest'(Wollheim 1987: 188). It is ro be distinguished from mere 'signatore'-those'characteristics of an arlist's work by reference to which we assign works tohim' (ibid. 197). He argul.'S that the traits which define a 'formed s~'le' may be'icry abstral.'"[ ones, of which \\e may not be immediately aware. And he addsthe interesting observation that commonly we start to notice these subtle char­acteristics only once we know that a p:lrticular p:linting actually is by a cert:linnamed painter, and not by another artist who happened to paint rather similarworks.

The 'generative' notion of lOt) Ic culminau..-s in what Wollheim terms 'sryle­descriptions', which arc dl.'SCriptions of the stylistic features of the work (Ithink he mcans by this, 'all the mature work') of an aOOt, at a high le\'el ofabstraction, so that one can appreciate the basis upon which this work is psy­chologically salient. He contrasts such a style-description to a 'srylistic descrip­tion' which is an account of the stylistic features ofany gi\'en work ofart, ratherthan of an artist.

Obviously, there is very little in Wollheim's programme for the productionof style--dcscriptions of individually identifiable painters with 'fonned' (dis­tinl.1:iyc) styles, that can be made usc of in the anthropology of art. Of course,anthropologist,s can study particular artists and their output, but if they doso, they are essenti:llly rceapirularing the programme of Western aesthetic artcriticism in some (relatively) exotic setting; they are not soh·ing the stylisticproblems POSl."<.! by the existence of collections of 'ethnological' art and theassociated cultural material, documents, ere. I-Iere, J am assuming thal what theart anthropologists deal with is unattributable, excepf to a particular culture,and that sorting out authorship is not the issue (but cf. Price 1989). On theother hand, there is something obviously exciting about Wollhcim's argumentthat stylistic analysis should be 'exphlnatory' and 'aT a high level ofabstraction'mther than tamely 'taxonomic'. The trouble is that on Wollheim's argulllent,it would seem that unless onc is dealing with artworks which manifest an indi­\'idunl style, the really interesting (i.e. non-taxonomic) questions cannot reallybe posed, let alone nnswered.

Is it feasible to n::I1'1",lIlge \Vollheim's programllle so that 'generativc' styleanalysis can be made applicable to 'gcneral' or 'historical' (i.e. cultural) style,and not exclusively LO indivi(lual style. as it seems to me he intends? I think this

158 S/yl<, alit! eill/ure S/yle 11IIt! Glil/ure 159

rather depends on the kind of material with which one is dealing. Wollheim isno doubt perfectly correct to believe that the important stylistic questionsrelating to Picasso, have to do with Picasso, not to his period, or to any of the'group styles' with which he was (closely or distantly) alliliated at differentphases of his career. And the same would apply to M.iehelangelo, or lngres, orany of the masters of the Western art tradition who possessed highly indi­vidual styles and who were, and ~lre, esteemed particularly on that basis. Butthis paradigm depends particularly on the social processes through which anartistic identity is achieved in the Western post-Renaissance art tradition, andcannot be applied to the kind of art I intend to discuss in this chapter.

Not only is the kind of art I am about to discuss defined with reference tocollectivities and their histories, not individuals, but it is also 'traditional' inthe sense that innovation was constrained within strict parameters of sLylisticcoherence. This is not to say that in these art-producing traditions innovationdid not occur; it did so, continuously. But it was not associated with artisticidentity, only with virtuosity. It is known that, for instance, individual tattooartists among the fvlaori achieved great personal fame (and charged higher feesaccordingly). They did so becausc their work instantiated, better than theircompetitors', what lvIaori collectively reg",ll"ded as excellence in the matter oftattooing-not because their tattoos were appreciated as distinctive produc­tions expressive of their artistic individuality. We may go to a dentist whom weconsider to be a supreme and indeed original exponent of fillings, crownings,and bridgework. All the same, we do not prize his work because it expresses hisindividuality, but simply because it is the hest of its kind available. It was thesame with Maori tattoo artists, and artists in traditional art-production systemsgenerally. Since in these systems there was no culturally recognized linkagebetween artistic excellence and the exprcssion of artistic individuality, andsince genres and morifs were subject to such stringent canons ofstylistic coher­ence, if is much more appropriate to treat 'collectivitics' rather than indi\'idualsas units of style, when dcaling with this kind of material, than it would be indiscussions of\Vestern art. Besides which, there is little alternative, given thenature of the artworks and documentation at our disposal.

Evcn so, is it really possible to provide 'generative' (abstract, high-level)style descriptions for 'ethnological' art, or must we remain forever in the shal­lows of taxonomy, detecting the signatures of this or that local style? Can oneachieve a style-description applicahle to the art of a 'culture' which has theS~lme kind of explanatory power that Wollheim e1aims for style-descriptionsapplicable to individual artists? One solution which immediately suggests itselfis 10 treat culturcs as analogous to individuals or persons, but on a differentscale, so to speak. nut this poses an immediate logical dilemma. 'Individuals' aredefined contrastively ,....ith 'collectivities', Picasso was a supreme individualistwho expressed his individuality in his arl:o I-Ie stands out against the backgroundof lesser artists lacking whM \Vollheim would recognize as 'formed' (personal)

styles, whosc artistic efforts lack this individualizing salience and arc conse­quently of lesser art-historical importance. One cannot treat 'cultural' art stvlesas magnified versions of individual art styles because, being 'collective' st}:les,they correspond to the 'b,lckground' against which 'style' (individual style,that is) creates its individuating effect. The notion of 'individual style' impli­citly depends on the existence of collective, undifferentiated, period stylesagainst which individuality emerges as 'figure' against 'ground'. One cannottreat collective styles as 'figural' in this sense, since there is no 'ground' to sctthcm ag.linst. Nor can one compare a 'cultural' style against another style asif cultures were individuals, in the way that one compares, say, Picasso andBraque. Picasso and Braque stand out as possessors of distinctive styles againstthe same (twentieth-cemury Western) background, whereas there is no com­mon background abrainst which one might compare, say, the art of the Fangand the Yolngu.

At the same time, it is hard to give up the idea that collective art styles havea kind of psychological saliency which is comparable to, though obviously notthe same as, the psychological saliency of individual styles. In so far as culturalanthropologists operate with a concept of style, this is a psychological (cognit­ive) concept, rather than a taxonomic one, or at least they would like it to be.Some anthropologists arc content to consider stylistic questions emirely withinthe taxonomic framework (i.c. identifying the provenance of museum speci­mens etc.) but most anthropologists associate 'style' with 'meaning' (Forge[(73). Anthropologists think of 'style' as the attributes ofartworks which asso­ciate those artworks with other culrural parameters, such as religious belief,kinship values, politiCill competition, etc. Since often the associations in ques­tion are communicated via the iconogr~lphy of im~lges (e.g. a carving is a carv­ing 'of' a totemic animal), anthropological notions of style are fatally emanglcdwith semiotic questions, to the detriment ofconceptual clarity. I will return tothe misalliance between rhe anthropology of art and semiotics later. For themoment, let us assume that iconography is not the relevant issue, but style, thatis, formal attributes of artworks. Can 'formal attributes of af[works' be associ­ated with other cultural parameters? If so, an 'anthropological' notion of stylewould focus on the psychological salience of artworks in directing attentiontowards cultural parameters. A 'cultural' style-description would be an abstractaccount of the attributes of artworks in the light of their capacity to thematizeand make cognitively salient essential cultural parameters.

8.2. Hanson on Style and Culture

This kind of progmmme has becn carried out by various anl'hropologists, thoughnot necessarily under the heading of 'style' analysis ('style' in anthropologicalwriting is oftell simply callcd 'aesthctics'). Very much ,l case ill point is the

160 .':;1)'/" ulld CuILuri' SI.I'It! and Gullure 161

article by Hanson (1983) in which he seeks 'homologous relations bctw(:cn

artistic forms and other structures or panerns of culture' (Hanson 1983: 79).Hanson's :micle is demtcd to Ytaori (traditional) art. First of all, he correcllymakes the distinction alluded to above between iconography and 51):lc; thoughhe speaks of 'aesthetic form' (following Archey 1965) rather than using theword 'style' itself. Nlost Maori art was not heavily in\ested with iconogI'Jphics)'mbolism, he So1YS (but cf. Neich 1996 and below, $1..'(;1. 9.6). On rhe otherhand, this an was Ilol 'meaningless' decoration because the formal characteris­tics of Maori art were such as to 'instantiate' underlying cultural patternsdetectable elsewhere in other ~'Iaori institutions. His main substantive pointwith regard 1O the art is the presence therein of mallY types of symmetry, lllter­nations, rotatiom, elc. (Han~on's llrride was published in the volume in whichWashburn published her initial paper on symmetry analysis as a method inarchaeological taxonomy, cf. Sects. 6.4-5, above.) llanson associates the pre­occupation with symmetry in Maori art with the Maori cultural emphasis onsequcnces of escalating reciprocal I.:ompetiti\'c exchange andlor warfare andrevenge. The idea is that from the Maori poinl of view, the presence of com­plex and rognitivdy inaccessible types of symmetry in, say, rafter-patterns, res­onated with the global, cxtr.t-artistic, pattern of i\'laori social life in whichsequences of reciprocal action, at various lncls and degrees of intensity, werepervasive. Moreo\Tr, just as reciprocity in sociallifc was never 'perfect' butwas always marginally unhal:lIlced, glving rise (0 the onward momentum ofcompetitive striving, so Maori hilateral symmetry is always marginally dis­turbed hy contradictory elements of wilful asymmetry. This tridiness in thejuxtaposition of symmetry-confirming and symmetry-disrupting elements co­incided with the Maori cultural presupposition that 'nmhing is what it seems'(Hanson 1983: 84).

Hanson's discussion is of great interest and is one of the most distinguishedcontrihutions to the literature of the anthropology of an. But there arc certainobvious objt:ctions that can be raised to his argument. For instance, both 'bilat­eral symmetry' and 'slighdy disturbed symmetry' arc features, if you lil;e, of a'uniycrsal' style as much as of 'the L\laori stylc'-that is to say, any patt~rn

llJha!son~r manifests symmetry because that is what a 'pattern' is. 'Dccor.lt­i'-e art' bv and br!!C consists of the applicuion of patterns to the surf::aces, , " ,of things, and unless these patterns were in vurious ways symmetricIl, theywould not be recognizable as patterns. As it is, there exists what amounts to a'universal aesthetic' of patterned surfaces; the same symmetry configur;ttions,if not the same motif", turn up ldl over the world, as \Vashhurn has amplydemonstrated in her more recent work (Washburn and Crowe 1992). So alsodocs the technique of disrupting the symmetry of patlerned surfau,,"S with asym­metric clements. Besidl."S which, asymmetries oftcn arise semi-inad\ertcntlybecause pattern-makers arc concerned only lO ma"e their patterns 'roughly'

symmetrical, not mathematically so. One cannot he sure therefore that Maoridisrupted symmetry is an intentional disruption of symmetry or merdy anum.:unccrn with eroc! symmetry.

So it (:Quld be objected that Hanson's argument about the pen'asi\'eness ofbilateral symmeuy in i\hori art is just a tautologous consequence of his deci­sion to look at 'fonnal' rather than iconographic aspects of the art. 'Formal'n:xluces to 'ornamental', 'ornamental' to 'pallcrned', and 'patterned' to'sym­metrical'. Meanwhile, is onc to assume that each and every cuhure that pro­duces ornamental (symmetrictl, patterned) :lrt has thc same 'cultural p:lt!ern'of escalating and disrupted reciprocity, and so forth? The Taj Jvlahal is emin­ently bilaterally symmetrical, :lnd slightly disruptcd (by the inscriptions)-burnobody has seen lit 10 compare Maori meeting houses with the Taj lvlahal,or to make inferences about reciprocity in Mugh:llindia on the basis of thisarchitcctur-.d preoccupation with symmetry (which extends to much other.\1ughal ornamental art).

I conclude from this rathet fundamental dilJicult} that Hanson's projcct issomewhat premature. The formal properties Hanson identifies in Maori art arcfar too commonly obsernblc in ornamental art of;l.lI kinds to serve the kind ofculturally diacritic role he proposes for them. He says .\1aori art provides akind of 'map' of Maori culture; but bow could this be demonstrated satisfac­torily if the dislincti\ e fealUrcs of this 'map' arc arbitrary in rclation 10 .\horiculture specifically, as they must he, if they arc also fou;ld in totally tliflcrentcultures, sueh as Mughal Inuia? The implication would seem to be that whatconnects Maori art to rhe rest of Maori culture is not 'formal' aesthetic orstylistic aTTributes such as symmetry, but the presence in the art of icono­graphic reference to Maori ancestors, divinities, etc.-precisely the position\\hich llanson set OUlIO question in the first place.

All is not lost perhaps. Hanson concludes his essay by suggesting that therelationship between Maori art\\orks and the rest of Maori culture is one of's~necdoche'; artworks arc parts of culture which recapiTUlate the whole. Itmust be admitted that in the light of the objections just raised, Hanson hasfailed to demonstrate the existcnce of thi... synecdochie relationship betweenartworks as parts and cuhures as wholes. Perhaps this is the rcsult of exccssi\etheoretical ambition; the allributes which (S.1)<) a mfter-pattern cJ.n possess,which can also be possessed by lltl entire culturc, would inevitably be highlyabstract olles. And for just this reason, it would be most unlikely that these veryabstract properties, such as 'symmetry', would be I!xclllsl<Jely found in the art­works of that cultme, and lhal culture alonc. The sc;u'ch for shared attributesbetween particular artworks :llld emire cultural systems is so grandiose thatwhatever is discovered is likely 10 prove fact itious. The idea of 'synecdoche' isnone rhe less intriguing, and seems to chime with our intuition that art st)'lesand cultures arc in some obscure \\a~ connected. Let us modify the proposed

162 Style liud Culture Style lil/d Cu/ture 163

synecdochic relation so as to make the whole enterprise less ambitious. Is anygiven artwork, in a given style (personal or collective), related by synecdoeh~

to till the artworks in that style.) The :mswer to this is surely 'yes'-because wehave unearthed one of the basic implications of the word 'style', namely, thatstyle lHtributes enable individual artworks 10 be subsumed into the class of art­works which share these particular attriol[[cs. Conscquently, any given artwork'exemplifies' the stylistic canons of the tradition of material culture ii'om whichit originates; it 'stands for' this style. This docs not immediately solve the prob­lem of style and culture, but it suggests a solution which will be outlined later.

8·3· S(Vle and Cogniti'ce Saliem)'

'fhe same considerations apply to Wollheim-type individual style-descriptions.An artwork which is being considered as exemplifying the individual style ofan artist, for instance, Pic:lsso, thematizes and makes s:llient Picasso's artisticintentions ~)ecause it enables connections to be drawn, on the basis of style,?etween thIS particular Picasso and other Picasso artworks. Even if there were,1Il fact, only one Picasso in existence (Cuernlm), the same would be true, onlvin this case stylistic analysis would be confined to generalizing abom the stvl~istic properties manifested in the different parts of this unique artwork-thehorse, the bull, the various figures, and so on. Unless it were possible to findconsistency in these part-to-part relationships in Gllcmica, the whl/le paintingwould be in no particular style, and would be 'indecipherable' (as \Vollheimput it) as a result. Such is manifestlv not the case. Cllert/ira is stvJisticallv deci­pherable and psychologically salien-t because part-TO-part relati~nships 'withinthe painting cohere with thc whole, and Gllernir{/ is an instance of 'Picasso'sstyle' (rather than 'Gucmira's style') because this painting is coherently relatedto other Picasso paintings, which are related to others still-so that ultimatelvall Pieassos arc intelligibly related to each. This is one of the reasons that w~esteem Picasso so much; however much his manner transforms itself, he main­tains a consistent stylistic identity, difficult though this is to pin down.

I hope it will be ,lcccpted, therefore, that the function of 'style' in associat­ing individual artworks with the totality of artworks 'in the same style' is notconfined to the 'taxonomic' domain, hut is equally pertinent with respect to theuse of 'style' as an explanatory concept with cognitive, as well as classificatoryimplications. It is unfortunately true that most discussions of 'style' focus onstyle in individwlI artworks (e.g. '\\fhat do we mean by the "style" of Cuernha,as opposed to the "content" or "significance"?" or, 'Is GlIerllim in the "cubist"style or not?'). I prefer to think of 'style' with reference not to individual art­works but in relation to the 'wholes' constituted by all the /lJorh in any givenstyle. 'Style' is whal enables any aTlwork to he referred to the whole(s), or

'larger unities' to which il belongs. However, this is nOll'O capitulate to meretaxonomy, because, according to my argument, the psyrlwlogiml stdiem)' ofart­TT'orks is a/imrlion ofthe stylistic relationship betmun any gJ1,xn art/IJOl'k and otherar/lPorks ill the same stV1e. Artworks do not do their cognitive work in isolation;they function because they co-operate synergically with one another, and thebasis of their synergic action is style. This is the basis of thc intuition we havcthat in some way stylistic affinity among works of art echoes the unity ofthought which binds mcmbers of social groups together; style is to artworkswh:lt group-identification is to social agents. BUI this is at best a cloudy notionuntil we have arrived at a much clearer understanding as to exactly what 'styl­istic affinity' means, particularly with reference to collective styles, such as theone observed by traditional Maori artists.

As we noted before, Hanson, in his over-ambitious attempt to find unify­ing patterns in _Maori art and Maori socio-cultural practices, only looked forabstract properties such as 'pervasi\'e bilateral symmetry' in the art. He docsnot consider the art in very much detail because such symmetries can be foundin almost every artefact the 11aori ever made, so it does not matter much whichspecific examples he chooses. Unfortunately, this unsystematic approach can­not exclude the possibility that the abstract stylistic features singled out aresuch as can be just as easily detected in lJon-Jdaori art, as is the case. How canthis problem be overcome? Clearly, in order to provide a style-description forMaori art which would be applicable to no other art style, the art must be sub­jected to a much more detailed analysis than the mere detection of a variety oftypes of symmetry and/or asymmetry. Attention has to be paid to the stylisticattributes which actually tell us that a particular artwork is a Maori artwork; forexample, the typical way in which the human hand is represented, the omni­present 'fern-shoot' motif, and so on. In other word,>, a formal analysis of theart has to be undertaken, quite withom reference to 'cultural patterns'. Onceformal analysis produces a style-description which captures axes of coherencein the Ivlaori artefacts whieh distinguish just those artefacts from :my otherculture's artefacts, then it might be feasible to attempt to align this specificstyle-description with the specifics of iVlaori culture. This Hanson cannot do,because he lacks the methodology to undertake such a formal analysis or to

identify the source of the coherence of the !\t\aori style in all its manifestations.

8+ Formal Allalysis and the Linguistic Chimera

'Formal analysis', alas, is nor a fashionable procedure in the anthropology ofart. The formal approach is doubly overshadowed in the age of interpretat­ive anthropology, in that it is either associated ,yith old-f;lshioned ethnology orwith the excesses of'el hnoscience'. The failure of 'structuralist' semiotic anlhro-

164 S(VI" untl Cui/urI' Style and C/llf1ju 165

pology in the I(nos has had a particularly pernicious impact on the status offormal analysis in rhe amhropology of art. At that period, it was eustOOlaf}'to discuss systems ofall kinds as 'languages'. Kinship was language-like, so wascookery (Uvi--5trauss 1970), so WOlS rcspccr-behfl\ iour (Goodeoough 1956), andso, neu..-..sarily, \\as art. Art was the (cultural) 'language of visual forms'. Thedominant position of the 'linguistic model' in cuhural analysis in the ethno­science period resuhed in the application, to \·isual 'language' of the linguisticmethod of decomposition into 'eonstiruenls' and the writing of (.l111stituenr'phrase-st'ruclure grammars', that is, sets of rules about how constituents couldbe combined int'O 'well-formed slrinb'1;', or :lcccptablc 'utterances'. Each <.:ulLurewas imagined to possess, nOl JUST 11 verbal hlllgwlge, hut various non-\'erbailanguages, one of which was the language of (:ll"tistic) form, or 'visual-esc'.The 'constit'Uents' of visual-ese were forms, typic;dly geometric forms suchas ovals, circles, lines, zigzags, ilnd so on. The phmse-structure rule...:; would,for instance, tell one how to combine lines and zig-L.ags so as to generate awell-formed utterance of 'poisonous snake' in, for instance, :\luba \'isual-ese(Faris r971: 103). A great deal of effort was la\'ished on \'isual grammars ofthis kind (Kom 1978) but il is fair to S:lV that the results were not commen­surate with the labour expended. As E. Gombrich remarked, in a reviewof Faris's '\'isual grammar' of Nuba body-painting, the end-product of thistype of analysis was only equivalent to a knitting-pattern, and no mote pro­ductive of insight. Whatever formal analysis might be good for, providinginstructions as to how to re-synthesize 'well-formed U(Lerances' in putative'visual langu;lges' is not one of them. \\tho wants to know how LO do Nubabody-paiming by being given step-by-stcp instructiol1S? Nobody bUl a N ulJabesides whil.:h, the instructions provided in pscudo-linguistic form arc inordin­ately complicated given the simplicity of the a(tual designs themselves; whichcan be easily (.'()pied by the most untalented draughtsman by a simple processof inspection and imitation.

The linguistic model foundcrs because therc is no hierarchy of'le\'e1s' in thevisual world corresponding to the multiplicity of levels in natuml languages,extending upwards from 'minimal constituents' (phonemes) to morphemes(words) to syntactic structures (words in phrases and sentences, expressingpropositions). Lines, circles. oV;lls, -,;ig7.ags, etc. are not '\'isual phonemes' (\'is­COles), The nOtion of"isual phonemes g'Jined currency partly as a resull of cer­tain well-known statements made by artists, notabl) cezanne (who was onlyrepeating the received wisdom of The art schools) th,11 visual forms could all bereduced l() cubes, cones, cylinders, etc. The idea thai forms can be assemhledfrom 'basic building blocks' definable in tcrms of p1:lllC or solid geometry is notonly ancient, but also pmctieal, to the extent thai this is often how artists actu­ally do apprO<lch the l'ask of dnllving, and indeed, 'b:ISic building blocks' of ageometric kind ma~ actually be implicated in the process of seeing, if l\ Iarr'stheor~; of vision is correct (1\larr 1982). But none oflhis be-Jrs an~ relation 10

the phoneme. Cubes, cylinders, cones, etc, stand lor, or represent, physical partsof objl.'t:ts, reduced to their geometrical bare bones, \\ hereas phonemes do norrepresent parts of what morphemes---<:ombinations of phonemes-represent.!\'lorphemes represent ((muftI! \\ ithin the linguistic code, but phonemes donot represent parts of these concepts.

Where geometric forms occur as graphic signs, they are meaningful in them­selves (a circle can represent nn eye) and where Ihey occur as parts (compon­ent,,) of other graphic signs (a eircle representing the iris of an eye, enclosedwilhin an O\'al) they stand for parts of whatever the gT'"Jphic sign as a whole rep­rescntS (the iris as part of the eye), By contrast, lhe lcrter 'd' in dog, does notstand for part of a dog, ilnd 'ogo' for lhe remainder of the dog. Any line whichis included in a dra wing of a dog, however, represents some pari of a dog. Thepart-whole relationships between the lines entering into the composition ofa grnphic representation, :md the representation ilS ;1 whole arc logically quitedistincl from the part-whole relationship bet ween the phonemes and mor­phemes. Consequently, the whole strategy of decomposing \·-isual presenta­lions into 'elements' or 'constituents' in the hope of writing 'yisual grammars'is misconceived.

8.5. Synecdoche: Axes oICoherellce ill S~}'listic Unities

The stylistic analysis offorms has to concern itsclfwith visual forms as wholes,rather than WiTh their geometrical constituents.' The aim of l"orrm11 analysis,as I sec it, is to pro\'ide a basis for the discussion of style. Style is a mcaning­less idca until il has been decided what the scope of the unit of style is, thatis to s,"lY, the works of an individual artist, a culture/period, ctc. as discussedpreviously.

In the present chapter, the 'larger unily' which will be analysed is the corpusof i\larquesan artefact... in particular the corpus a\"3ilable to Karl \on denSteinen, who published a lavishly illustrated threc-mlume tn.-atise on .\Iar­quesan material culture in t925 8, ha\'ing visited the islands himself in 18c)8in order to make his own collections, document them ethnographically, andrecord the tilfloo designs whieh will be discussed below. The aim of formalstylistic analysis is TO show how each particular item in the corpus is conncctedto the corpus :IS a whole. In order to do this, there is IlO nced to make any useat all of :lny linguistic analogy.

, To the C'lcnt that there is allY mileagc at all in thc id.,.a of'"su"l syntax· it concernS principLes ofarranl!C1t1cnl or ~V"'posilion,such as th~ "hich are n,anifest~~l ,n thc byoUl of Yotngu barl-l'ainting.as described and analysed by \torph~ ('99t, '99~)- This t,arlicular appronch will r>(){ be I'u~ued here.ho",,'cr

166 S~)llf allli GU/lUre Style 1111I1 Culwrc 167

However, there is:l sense in which formal analysis is imrinsically scmioric,if not linguistic. This is the sense in which any pan 'stanus for' the whole ofwhich it is a part (synecdoche; cr. above). Any example of IVlarquesan art is, toa greaer or lesser extent, representative of the corpus to which it can bereferred, and does indeed rcprl..'scnt iL 'Representing' in this sense is clearly asemiotic relation, in which the object is a sign, and the (:orpus of SI)lislicallyrel:ned objects from which it is dr:I\\n, is what is signified thereby. This kind of'sym.'Cdoche' is not "ery scm:lnlically infonnau,-e, since c\ cry work in the corpussignifies almost exactly the same thing, namely all the other ones except itself.

Howcnr, synecdoche does suggest another useful mctaphor for thinkingabout sl}"le. There is a kind of image known as a hulogt"'Jm (made by photo-.gt"Jphing thc interference fringes rent:<:ted off objects illuminated by a coher­ent laser lighr-source) which has the curious property that any part of ahologt"Jphic image contains an attenuated version of the information containedin the hologram as a totality. From, say, the hottom left-hand corner of ahologram, it is possible to reconstitute (by shining tlser light onto the etchedholographic plate) the image of the original object, though t:\thcr blurred. Stylein art is like this in thc sense that from one item in the corpus (or a selection ofthem) it is possible to reconstruct the others to at least some degree. Of course,one docs not know which features of a given objcct .1rc the stylistically sig­nificant or informative ones except with reference to the corpus as a whole,so the analogy is inexact. Eaeh objeer, seen in the light of all the orhers in itscorpus, appe-.m; as a micr(x:osm of the corpus because our perception of it isinfonned by our knowledgc of the macrocosm of \\ hich it is a fragment.

The position is akin to that of the palacontulogist, who has to reconstructan extinct spl.-'Cies on rhe basis of a single bone, as Baron Cu\·ier was famouslyable t.o do. em'ier could rt..'Construct an entire animal on the basis of say, asingle femur, because he "'as familiar, not just with the relationship betweenthe femur and adjacent bones in the skeleton, but with the 'relations bcnveenrelarions' in the skeletal architecture of the wrtebrate genus under considera­tion. He could, so to speak, sec a femur as a rransformed \·ersion of the femurofanimal.. whose morphology was known, and by applying the same consistentseries of rransformations seen in the isolated femur to all thc other bones ofthe exrinct animal's skeleton, he could reconstruct ;ts form. He could, in otherwords, infer rhe 'style' of the fcmur by comparison with the femurs of relatedspecie..<;, and he could apply this 'style transformation' to the missing hones ofthe rest of the skeleton. The relatiun between the single bone and the skeletonis 'holographic' nOI because the isohHed bone oUlwardly resembles;n form theolher, missing ones, but because the comparison of rehlled forms suggests atransformational rule for 'generating' missing bOlles from known ones.

Rather than think of formal analysis as an adjunct to scmiotics, it is morehelpful to imagine it as 'morphology' applied to \"isllal objects. The aim is to

derive, by the comparison of related forms, a serics of transtormations throughwhich given artworks or artefacts can Ix: converted into other ones. The aim isnot, of course, 1.0 appl) thesc transformations so as to 'pnKluee' new (hypo­thetical) artworks in thc way a maker of forgcries does. The making of forgcriesrequires no intellcctual elaboration uf srylistic theary, being a quite intuiti\'eprocess. The purpose of formal analysis is to identify a;r~J of('oh~mlre withinrhe corpus of works brought together for the purpose of sl}·listic anal)sis. Oncethese axes ofstylistic coherence hne been idcmified, it then becomes possibleto understand lhe cogniti\'e significance of a 'cultut"J.I' style in rendering fea­tures of thc culture cogniti\·e1y salient.

In a sense, cogniti\·c sensitivity to 'axes tlf coherence' and the application oftransfonnational rules is built into the process of seeing itself. An) familiarobjt..-ct, such as ,I pair of scissors, is intrinsically capable of casting infinitelymany highly dissimilar images on the rerina, which the brain has to interpret.Looked at end-an, from the handle end, :1 p:lir of scissors may appear as it flatbar of metal, curiously shaded in the middle. resembling thc 'canonical' pict Lirewe have of scissors in no obvious respect (i.e. thc picture of scissurs we mightinsert in an illl1Slr,lted dictionary). nut our visual system, according to Gibson([986: 73-5), is tUlled to detecting solid objects because they map onto otheravailable views of the same object that can be obtained by shifting the point ofobservation. An object of cumplicated contours is seen as 'structurally invariantunder transformation'. In the same way, style analysis looks for the 'structuralim-arianls under transformarion' which define objccts of it higher hierarchicalorder than the isolated object, such as a pair of scissors. Thus, 'ailihe works ofRembrandt' could be considered, not a collcction of separate objecrs, but justOl/e object with m:m) parts distributed in many different places. A single objectwhich is a set of subordinate objects is not hard to imagine; a t\\ehe-piecedinner set is one such object, a chess-set another. If we take Rembrandt's'complete \\orks' as such a multiple, or 'distributed' object, then it will be secnrhat the praccs.o; ofdetcrmining Rembrandt's 'st}lc' is prl.-'Cisely the same as lheprocess of dcremlining the 'im·arianls under tT:Jnsformation' of it Gibsonianobjet.."t, such as a pair of seissors. What is 'ill\ariant under transfurmation'is what links anyone Rembrandt work to all the othcrs, bur unless we comsee what these invariants of structure are, \IC arc not seeing this, particularRembrandt, 'as a Rembrandt', that is, as.1 component of an U'uvrc. 'Educatedperception' of a Rembrandt amounts to nothing else. Stylistic perception, 1argue, is the perceptual mode with which we deal with multiple or distributedobjects of this kind. As such, stylistic theory is just an extension of the theoryof perception itself. llowever, it is a \·ery spcci.llized task to conceptualize wh.llthe 'in\;ariants under transformation' in Rcmbrandt's \\"orks are, and I <10not propose to undertake it-though [ bdic\·c that lhis would amount toa Wollheim-type 'style-description'. It is much c:lsier to operationalizc this

16~ Style and CuI/urI' SI),It- III/d Clllllirt 169

notion of stylistic l.:ohl.:n;m;l.: with respect to J\'1arquesan art, because, as I willshortlv lh:monslrare, Marquesan art is particularly suscept'ible to :1l1:11ysis interms-of visually salienr morphological transformations. Rcmbmmlt's sryle ishidden in the subtleties of his handling, whereas Marques.1n style or such ofit as we now have access [()-Gin be disinterred with paper and pencil.

8.6. Tlte MarfjlleSlIlI Corpus

The artefacts in the .I\'larquesan corpus assemhled hy Steinen (t925) arc par­ticularly suited to morphologil..-al study, because the~" form a \ery coherenl set.The corpus (;onsists of a number of categories of artefact..., of which the mostimportant arc (i) tattooing motif... (ii) similar d«oration on artefacts such asflutes. (iii) incised work on the flat, that is, Clrving in very low relief. and (i\')three-dimensional can'ing, some of it monumental but much of it on a small,and e\en minialure ~-ale. Of these categories, t;lUooing motifs arc I..-errainlythe most important and best documented, though during the early period ofEuropean contact (from the late eighteenth to the middle of the nineteenthcentunr) the Marquesans were prolific producers and consumers ofartefact... ofall kinds. Marquesan society was orientated towards magnificence and displayto a gr<:ater extent than any other Polynesian society, with the possible excep­tion of the "laori of ~ew Zealand (SI..'C Gcll 1l)!')3; Thomas 1990). I will saymore about the social, political, and ideological charach..-ristics of Marqucsansociety at a later stage. For the prescnt I am concerned only with Striclly for­mal aspects of Marquesan art, and it is sufficient to note only that almost all!\'Iarqucsan :Jrl was attached to the human hody (e.g. tattooing, adornmenr)..\10reover, the art that was not inrrinsiCllly part of the human body (e.g.weapons, canoes, furnishings of houses, etc.) \\as conceptually lrcatcd as if itwere. Thus, a chicf's l.:anoc was pan of his body, had a personal name whichwas one of his own set of names, if injurr was done to it, injury was done tohim, and so on. Conscquendy, it is rcasonahle to commence the study ofMarquesan an, as Steinen did, by considering tattoo art, in that this art reallysets the pattern for all the others.'

The following analysis owes much more to Steinen than the data it uses.Steinen does nOl cile the art-historical literature of motif-analysis of his period,the work of Riegcl, Goodyear, and Wickhon', which has been excavaTed for usby Gombrich (1984: t80-zoo), but he probably knew ufit, since these wO.l"kswere well known to art experts of Steinen's generation. Steinen's theol'etleal

, Of coursc. Tre~ting hody-art as pr;Tl1~r)", and such arts as carYing:.amI "nST;1 "~I ".11 ~_, ~cC,.'"tI:I:)", is .uneccemric pmeetlure frum Ihe SI~nJ)X1int ur Weslern uri thcory. whtcll concern, Itscll prun~rll~' wnhpainlinlls ~11.t sculptures ";CII ed "S independent :ICS! hClic e~lilic, r.lthcr th"n ~s Il~c mere. apPllrle,~:,nc('sof li\"ing: \)(,;ngs ~nd social "gems, Tattooing in Ihc WCS! I., a '-ery second-r:ltc furm o! orI, hlll tI' theJ\larqucsan COlllC.~tlh;~ ,,"'s nOl so.

framework (like Riegel's) is evolUTionary; he is concerned to chart the historyof motifs from relatively 'realistic' versions documented in the earliest imagesof Marquesan tattoning (especially from Langsdorff's publication of 1804(1813-'4») to the very much more abstract versions he himself documented inthe ,8ljos. Hut evolution:lry and l.:ompannive speculation hardly predominates;he also devotes a great deal of .mention to synchronic \"ariation in motifs, andparticularly what one can only descrihe as derivational or transformational rela­tionships between motif... and forms. These he elucidates in a series of analyt­ical figures which arc frequently both ingenious and reveiaLOry. Steinen, whomade all his o\\n illustr.ltions, was a graphic artist of remarkable ability, andhis three~\'olume work on the Marquesas is arguabl~' the most aestheticallyaccomplished work e\er LO have been published in the discipline of anrhropo­109)" (the design and layout of the edition is a bibliophile's dre-dm). I hope thatthe following pages, which, I emphasize, rely \'ery extensively on Steincn'soriginal work, will encourage wider appreciation ofhis contribution. He was, Ithink, a forerunner ofstructuralism in the domain of material culture, like hiscompatriot Goethe in the domain of plant morphology. The comparison is apt,in the sense that just as Goethe imagined that all plants wuld be seen as mod­ified versions ofan 'Ur-plant' (anticipating an imponant element in eyolution­ary thought) so Steinen sees !\larqucsan motifs and artefacLS as transformationsof a series of Ur-motifs, the I!tun (godling), the 'face' motif, the 'chiroid' (half­oval), the 'woYen' motif, and so on. Each of these can be transformed into theothers, as he dcmonstroltes. Steinen's preoccupation with the 'derivation' or'etymology' of motifs is embedded in his nineteenth-cenrury e\-olutionary(de\olutionary) mind-set, and to this ext.ent I shall not follow him. On thecontrary, I would argue that fhe mOfivic tr.lIlsformations and deri\'ations hedemonstrates cannot he placed in :111 evolutionary framework, since the datalack sufficient time-depth; all the reliable \'isual documentation of tattooingcomes from the twenry years spanning the turn of rhe present ccntury, whileearly reprc.'ientations of Marquesan tattooing, including LangsdorfT's, are ofquestionable accumcv. Even the carvcd artefacLS Steinen discusses arc almostentirely of (probable) nineteenth-century date. I hope to show that the styl­istic features of Marquesan an arc lX:Sf explained in terms of synchronicideological and cultural forces that Stcinen tended to overlook, because of hise\'olutionary pn:occupations.

In whatever wa\' the transformational or derivational relationships in Mar­ques~lIl art are 10 be expbined, the first necessity is to dcmonstratc thcir exist­ence, aOlI here I follow in Steinen's footsteps.

The l'vIarquesan art style is nn :lsscmhlage of artistic practices demonstratinga coherence which is visually immediately apparent, bUl which only carefulmotif-analysis C;lll fully reveal. For 1he purposes of this analysis 'relationships'between motifs are e.\pressed by procedures which permit one to turn any

liD iiI

8.7. 'lite Tllble ofErua Motifs

FIG. 11.7/1. [/UII: the c1;\Ssifical0~· scheme of (he domin~nl :\bnluesan motif rtUIl d.,,-elop.."d b~

Karl ,on den Sldn"n

E r

11: 1Cmt.X:-:~n=C\'\~xo.b C defghikl

~ :li~ ~ ~1JJLlJJ.M1J. b1L r'i nIDID.

~---~ ~--~---i =u H [ K 1. M

:n:&1Q~~~~&~PWN~W¢

r S !'------- .. ~-- 1: ------------.1

~~~;i;~~'U. V (.oJ X Y ~

For sll'listie relutions bClWeen ~II/II motifs in this table and other i\larql1es:m tl101ifs sec Fi!t".ll,7!2for lh~ trmsformation of rill II A 10 ~[UII G; Fi;:. 8.713 ror ,'[lIdS I., ,\t. and 1\, Fi;:. S.Sjl for (illil 0and f",p~ Ulu"l/f llloti~~: Fig". 8.'.Ils lor mill N and ~'fll Il\(Hifs; Fig. 8.lOjl for ,'[1111 I" and thcmilia fWIl/Illl1IJtif; Fig. B.lll;! fOl" dl/III and d\l' '''''111/11 panern. Sm"cr: Von ,kn Stcil1etl. i, tS;!,

illus. 100

reduplicate the person by appending additional, subsidiary 'persons' 10 il ingraphic form (cf. abO\e on fractal personhooJ, 7,10-11).

The mOlif wilh which to commence therefore is ell/fl, the generic \\ord forany kind of goJling.

given motif (or form) into rmother motif(or form) found in the corpus. By um­{rast to the motif-analyses of '9705 ethnosciel1l:e or 'visual language' analyses,the aim is not to 'gener,He' moti[<; from their elements. but to proceed. \-iaspl.'cificd steps, from one mOlif tn another; that is, the analysis presupposesonly one level, that of the motif (or tann). .\I01i[<; arc recognizable both beeausethey recur in identical form repe-J.tedl} in the (,1)rpus (as well as modified forms):md because they have documcnRx! motif-names, some of which are cullurall}rc\·ealing, though by no m(,-ans all of them. I will not discuss the implil.:ationsof motif-names at this stage, but I will use the .\ larquesan names to identif}motifs as a matter of cnn'·eniencc.

The types of procedures for tran.<;fonning one motif in lhe i\tarquesan cor­pus into anorher motif C'J.n he briefly enumerated at this stagc. They are:

These procedures (,-an be applied both to motifs as ;1 whole, or to parts offfimifs. Thus, in order to lr.ll1sfoml a motif into another motif it can be splitapart, and just parI of it may be subjected to a motion in the plane, or a changein relative proportions Ctc.

Thl~ outcome of the analysis of forms along th(,'SC lines is the creation of anetwork of transfomlation:l.l rel<ltionships among motifs. This nemor!.: of rcla­tions has no centre in that therc is no one motif which L-an be rq~ardcd as theapical ancestor from which all the rest arc deri\'at;ve. TbL'Qrcrically, all the rela­rions des<..Tibcd in this nctwork are symmetrical; if mot'if A can be deri\·ed frommotif B, then motifR L-an be deri\ed from A, there being no imputation ofC\'o­

lutionary priori~- either way. Howc\-cr. the analysis has ru begin somewhere,and it is certainl\' convenient to commence with a motif which has as muchvisual saliency as· possible. Srcinen's discussion of tattoo motifs is divided infOthree sections; l. 'Plcctogenc Muster' patterns which relate fa plaiting, mat­ting, basketry, etc.; II. 'Tikigcne J\luster'-patrerns which contain images ofthe human both; and 111. 'Gesichtsmotiv'-motifs which sholl' or arc derivedfrom the face. I' will le,we 'Plcete>genc Muster' aside for the moment and willconcentrate on II and 111, which comprise the most eUlllplic;lted and visuallyinteresting motifs. Of these twO, it is yon den Stcincn's 'Tikigcne Muster'which provide the most natur,Jl point of departure, As I noted enrlier, l\tlar­qucsan art was either placed directly on the human person, ~)~. on ~rtefacts w~lich

were tre-.ned as persons, The 'Tikigene .\!ustcr' ,Ire muuts which cssentmll~

(;) ,;gid mo';on' in ,he pi,", (plone rmn.,fnm",;on" rmn;l"ion, co"";nn,ref1(,'Ction, glide refleclion; cf. Washburn and Crowe 1988);

(ii) l"Oordinate transformations (modelled on those of D'Arcy Thompson(19(JI: ch. 9 (rtJI7)), i.e. changes in the proportions of motifs;

(iii) hierarehizationfdecomposition (one motif becomes a component ofanother motif, Ot a component of a motif becomes a motif);

(iv) transformations of dimensionality, i.e. a tw(}-dimensional motif is tr:ms­formed into a three-dimensional motif and vice \'et'S:l.

172 Style lind Cullure173

__ N

FIG_ S.iIJ. Rcncetionand rolation in therormalion or rillil 1., M,and N In Fig. S.ill

,...k'

....k'

tm. 8.i/2. Plane rotallon in Ihe formation of Steincn's two dOlSSC$ of (fila

Stcincn's method of rnotifanalysis is displayed in his table of etlla motifs whichI reproduce in Fig. 8.i1I. I shall usc this table as a point of departure forexploring the entire range of Marqucsan motifs. In this table Stcincn distin­guishes seven types of (tua motifs (in rows) which he groups into two maindasscs; 'standing' Cilia and 'sitting' elml. :'-Jo iconographil: significance shouldbe anached to this distinction, which is entirely convention"l. A 'standing' t:flla

is one whose lower limbs form a convex curve relative to the botly vertical, a'siuing' (ilia is one whose lower limbs form a com.-:lVC curve relative to the bodyvertical: see Fig. 8.7/Z. This fig. contrasts A and G on the table of ttua.

As will be seen from Fig. 8.7/2, the 'standing/sining' comrast in duo motifscan be produced by a rigid motion in the plane aJTecting L":l.eh of the lower limbsof the ~tua, namely, a rotarion through 90 degrees. This is a typical instanceof the ','isual logic' of .Marqucsan art, and is the first of many examples of asimilar kind that I will discuss.

As inspection of the table of ~/lIa shows, types of standing or sitting ~tua

show a high degree of \'anability. The uppermost row arc examples mken fromrelief L":l.rving (on stone, bone, or in the case of E an eluo silhouette in tortoise­shell). The next two rows of st:lnding e!lUl are very reduced forms, found intauooing or in engraving on bamboo (whose motif repertoire is identical to tat­tooing). I will deal with these reduced forms later, since, as will be seen, theymerge into Steinen's other major category of Plectogene Muster, 'plaited orwoven' motifs. It is on the 'sirting ellla' in the fourth and fifth rows thatinitial1y [ shall concentrate, especially L, Nt, N, and 0. L is a linear version ofG, used as a tattoo motif, and Nt is its coumerpart, produced by subjecting the'hands' to rcHeclion ill the vertical plane (Ot alternately, exchanging the leftfor the right hand and vice versa). To produce the 'squared-otT spiral' arlllform of N, :mother reflection (in the horizontal plane) is all that is necessary(Fig. 8.7/3). The 'legs' of 1'\ arc produced by subjecting the 'arms' so pro~luced,

to a reflection in the horizontal plane and a rotation of 90 degrees antlclock­wise-this leg-form can also be seen as a reflected, modified version of the basic'sitting' leg-form seen in G, the starting-point.

A slightly different leg-form is seen in 0, which is otherwise a close rclatiyeofN. This leg-form is the same as G but minus the nick at the base indicatingthe boundary between the thigh and calf of the leg in the squatting position.'rhis line, duplicatL'<.1, produces the concertina-like leg-forms ofl and K, whichlead lOwards other motifs I \\iII mention later. Bur the leg-form ofO (which isl/-t of an ellipse) eliminates this line and can be treared as the origin of a verylarge class of semicircular and paraboloidal motifs.

Motif 0, in male hJttooing, is anrhropomorphic, since, like L, M, and N itis a form ofKena. a mythological hero connectl..'<.1 wilh tattooing, who is shownas a sitting etua. But according to Steinen'5 information, the identical cflla formin female tattooing is named Kea, which he identifies as referring to the tOr­toise. These details of nomenclature would be beside the point in a formal ana­lysis such as this but for one thing, All members of the tortoise/turtle familywere considered sacred and divine by the i\-1arqucsans, so there is nothingsurprising about species of lest/ufo appearing as sacred eflla in tattoo motifs.The point to note though, is thal the canonical view of a tortoise is in thehorizontal pbne, para Ilel with the earth, whereas Ihe canonical view of a sitting:efua is vertical relative to the plane of the earth. So there is another kind of'geometric' transform,llion here depem1cnt on whether the image is viewedas a section or a plan: a sitting etl/a becomes, when the implied orientationof the observer and the I!ftltI are shifted fro111 the \'errical to the horizontal, a

174 S~vle lind Cullure Style lind Gullure 175

FIG. 8.714-. Gtomdri~ orienlation affecting thc change of motifs: from dlUl to tht femalttattooing tortoi,e motif kCI1

vertical

silting elua 0 in

Fig 8.7/1: Kerla

horizontal

tortoise: kea

which are st:mdard hope vehine motifs), These smooth transitions indicate anevolutionary pathway, but, as I noted earlier, I am not so interested in evolu­tionary relationships as in transformations which can be carried out on onefigure so as 10 convert it directly into another, without intermediary steps.Consequently, from my poim of view, it is more useful to think of the rela­tion between the sining [/,/11 and hope vehine as shown on Fig. 8.8/2. ln thisfigure, it is shown that the cilia can be transformed inlO hope vehine by split­ting an cIlia of the form seen in Fig. 8.8/1 vertically, and rotating one half ofit through rHo degrees before rejoining the two halves. This produces a hope'I,'ehine-like Fig. 8,8/1, row 2, ii. J

tortoise-divinity viewed from above, without any actual change in the mor­phology of the motif whatsoever (Fig. 8.7/4). The horizont.al (tonoise) form ofthe sining eilla is highly productive of further motifs (other forms of ken andkake which I will discuss in due course).

8.8. From Hope Vehine to Vai 0 Kena Siting ell!<> --FIG. 8.8{2. From siUing; e/Utl (Fig. R8/1,row I, no. iii) to hopc "Chllie (roll' 2, ii)

~~

~5t1~~~~51ill~~~~N~

FIG. 8.8/1. Etua 0 in Fig. 8.7/1 sliJing: by tlegrees to bCCOlllt the h(ljJC ,;chine cla,s 01 motilS.SOUfCC: Von den Stcinen, i. 16." illus. II I

Taking 0 as the starting-point, I shall next consider a very prominent motifwhich goes under the name of fwpe vehllle (literally 'buttocks-women', which Ibelieve refers to back-to-back siamese-twin fem:lle divinities; cf. Gell 199]:(93). This motif can be derived in a number of ways, but most readily f;olTl aform of ewa such as 0 or 'r' in the next row in Fig. 8.7/1. In a separate figure,Steinen sholVs the derivation of hope uchine very perspicuously by the com­parison of a series of related forms, taken frOI~ tattooi~g (Fig. ~.8/r) . .Thisfigure shows the sitting e/lUl sliding, by degrees, mto the form of dmmet~lcallyopposed quarter-ovals which is distin~tive of hope ve~ine \:ia a ~edu~tlOn ofone 'leg' of the eflltl, and a eompensatmg transformatIOn of the arm on theopposite side into a quarter-oval form (ef. row T, ii, iv, v, and row 2, ii, iii,

dI'o'ido ko"ht.....,.....-d relate \00'

bel"'" rojoirtr,g

Hope ,-'ehine, illlllrn, can be subjected to transformation so as to produce fur­ther motifs. The first of these I shall consider is the one known as utii 0 Kena,the bath of Kena, the tattooing hero mentioned earlier, whose mythology Jhave discussed at length elsewhere (Gell 199T 186-8). 'rhere arc three docu­mented forms of this motif, each of which is derived from hope vehine in aslightly different way. The most elementary form of t;ai 0 Kena consists of110 more than a hope 'Veh/lle and its reflection. t\ different, more complex, formof VIIi 0 KCl/a is produced :lS shown in Fig. 8.8/3. In this figure, a hope vehilleis subjected to reflection and translation so as to form a refJected duplicateadjacent to the original, to which it is joined. The next step is 10 'stretch' eachof these hope uhille motifs horizoIltally so that they overlap in the middleand partially occlude e:lch other. This produces a form of STanding ell/a motif,in which the opposed quarter-ovals of the double hope vehille become ':lrms'and the conjoined quarter-O\'als become 'legs'. The merging between thisand the ClIIII motif is completed by the addition of vestigial 'head' :lnd body e1e~

ments.

) tkadcrs of Ihis \I ho happen w bc bmiliar willl Iht I; pe of CIJlling, in"erling, and pa'ting opera­tions to Id\ich Le-i-Slrau" subj,'cls ml"ths in order 10 demnns,rale Ihe affinitics betwecn myths"hi~h do nUl. ,m the wrf:,ce, "Ppea,. to be al ,,11 similar. "ill appreciale lhe significance of thi. IYpC ofprocedure.

176 Slyle «lid Gu!Jurr S~l'le I//Id ClIlllIft 177

rll;. 8.')/"1. Kilt" from."ft uhl1l(: duplication,rotation, ~nd rcllcetion_After Willoll'dcan Handy,platc XWI

In the case of 1:«1'( motifs, the coordinates arc much more distoned though.The kaKe shown in Fig. 8.9/2 (from \Villowdcan Handy, plate XXVI) consistsof duplit'"3ted hope vehillc, one of which is translated and joined to the first.However, the symmetry of lhe two hopt' velJilll! is disguised by changes inrelative proportions; the top left and bottom right isolated quarter....O\alsare reduced, while the two quarlcr-orals which join up are expanded (Fig.8·9/2 ).

fiG. 8.8/3. From hOM ullin(to t'/II 0 Kt'IIlI

The most complex \'cmon nf raj0 Kma is also based on IIOIM ;:~"i,,~ (Fig. 8.813,III). To produce this ,,'0; 0 Keno the hopt IX"i"r motif is first of all duplic:uedand rOlated 90 degrees. The resulting superimposed "opr crhine mOlif... arc thendrawn apart !>O as to produce a square motif four times larger, in which eachquarlL'T-oval O<.'CUpies onc comer. Between these qU3rtcr-ovals, the originalblack conca\ e lens shape secn in hopt ;;rhini' (now duplicated) appears as a blackX joining them diagonally. Finally, the small square forms un the 101' and thebottom of hope venint, not included within the quarter-o\'als, 3rc extendeddown to become the central axis of the motif, represeming Kena in his hath.

8.9. From Hope Vchine to Kake/Kea

Then thc entire figure so fonned is stretched and bent so that the smaller quartcr­O\-al5 become the head/tail of a 'podoid' or 'footed creature' in Stcinen's ter­minology. Steinen also has a different. form of kake, which is derived directlyfrom a sitting etlla, rather than from hope uhilU (Fig. 8.913). To produce a Steinc.ntype II flake, a sitting etua like 0 in the (able of"Iua is duplicated, and the dupll­catc is joined to lhe original aflcr 180 degrees rotation, producing a differentkind of 'podoid'. This is then stretched and twisted, as with the type I kake,Fig. 8.9/1, ttl produce the char.:acterisri(''"3lly bent kake motif. There is a subtledifference in S\'mmetry between type J and type II kake motifs, ncn when theyappear to be double-headed hope vehine motifs, depending on whether thedupli(''"3te has been rotaled (as in type 11) or not, lype I (Steinen, iii. 162, iIlus.110, \'s. Handy, pI. XXVI). Kal'e can be, and commonly are, drawn in pairs, onerotated ISO degrees rclalhc to lhe other, so as to fill a reclangular space.

FIG. 8.9/1. Coordin;lIC U'Jns(Qrm~llon

"f a hQpe t'l'hin( motif resulling in 10,,1:((Iype I) from \Villo\\"de~n Haml}.S&ura: \Vi1Io\\"dc~n Handy, pJaIC XXI'I

Next, I shnl1 turn to the second category of motifs which c:ln bc dcrived fromhope vehille, thc fUifie ('climbing') motifs, which merge into kc(/, :lIld :\I'e some­times known by that name. The kal,c mOlif is a stretched ;1I1d tWIsted hopcr,;chilll! (Fig. 8.91 I), :In installcc of 'coordinate tr:ln~form;~tion',in th: 1ll:111ner ofD'Arc\' Thompson, l".uhcr th:lll symmetry transformatIOns a In Wasl~burn ofthe tYI)C we ha\'e hcen considering hitherto-though. some 'srrelch1l1g' wasinvolved in the IIfJPf vehill(~ vai (} Ko/(/ example r discussed a mOIllCIl! ago.

FIG. 8.9/.1. CoonlinalctTansf"nnation with romtion:rrom sitting <'111// to j'l/h

(lype 11) from Sl~incn.

S,,"n,,: Von den Stcincn, i.16l, illus. 110 (ahow); i.t60/, iIlus. 110/ (below)

179

Finally, before le.wing 1:(11:1:, I m3} mention a unique design recorded byH3ndy (Fig. 8.9/-+; I hlndy, pI. xv) which is the 'girdle' tatlooed ahove the but­tocks and OYer the sacr.11 area of a female chief. Sleinen n:conled nothing likeit but that may be because as a male ethnographer, he was not in a position to

persuade this particular female chief to re\"Cal her tanoo. The large motifs tothe left and right are (single-headed) kaJu (Handy herself wriu.'s 'ka'ake', whichI think is mistaken). Such large 'loin arches' were a feature of many Polynesianunooing traditions (e.g. the Sociely Islands; see Gelll993). What is sl)"listie­ally interesting though is that here we have an instance of hierarchization,which I ha\e hitherto not had 3n} oceasion to mention. All the designs I ha\'ediscusst..'d so far are quite small, but here we see a double pair of (mirror-image)l'al"r (four in all) as m3jor compositional ek'lTlents, subsuming many subsidiary'reduced' ctllo figures (PollII andfimatw to be precise). Furthermore, the wholeassemblage may be read as a '(.1ce' motif on a \"Cry large sC'J.le, though I shallreturn to this possibility later, when I have deah with The C/tw ---? '£.1cefeye'motif transition. Furthcr instances of the inclusion of motif.<; within othermotifs will crop up ag;lin, but this is a striking example.

•

FIG. 8.914. Thc ehicfcss's ginl1e with lyre III J:llk~_ Sec Section g, 11 below for facc motif,. S9uru:

\Villowdcan I landy, platl: xv

It will have becn nOledthal the heads of the 'podoids' in Fig. 8.91-+, resembleneither our type l or lype II kalu very closely. They rescmble, in fact arc iden­tical to the small ell/a immcdialelr adjacem to them, which are instances oflhe ellj~ motif Pullu (a mythological hero) here depicted with panicularlylarge hands (or forefcct) which these kola: also havc. They arc similar in thisrespect to another (;ategory ofelongated, but nOllwisted 'podoids' categorized,

FIG.8,.Hj· K<'IJ (lO"oisc) fo:nule UIlOOmulifs m:orded b~' Stcinen. Suurrr.

\'on den SlcilKTI, ii. 205. iUus. l06

once agnin, as kt:a (resrudinares). These are pod()id~ which are produced by sel­ting etua motifs end to end in opposed pairs (i.e. one rohltcd 90 degrees; sceFig. 8.915). Such kelt are especiallv common in It.mel tattoos from which thcseexamples arc drawn (Steincn, i. [66, ilIus. I19). A large selection arc foundas decoration applied to a soup-bowl (terrinc) which is itself the shape of atortoisc, where they arc arrayed like the platcs on a tot"loise's shcl~ tortoiseof tortoises: Fig. 8.9/6. In Ihese kt:a, the eight legs ma~ be turned into squared­off meanders, rather than the spirals seen in the original ellia motif... (i.e. ;.Jon row.) of the lableof eilla, Fig. 8.7/1).• In so far as these 'tortoise' motif... ere-

FIG. 8.9/6, ablll·~. ..\ lortoise of tortoises: t'M motifs on 10rlOisc·shaped ltrrinc. SOflra: V"l1 Jen Stdntn, ii. lOS. illus. 206

1'1(;. S.')I" leji. 'Re:llistic' ~"'II lortoise taTlOO rttordcd by Lungsdorffin [g04. SOU",,: Von <kn Sleintn. i. '91, illus, q8b

, An iOlcrc,ting questi"" i~ rJis.cd b~ the fael IhJ' ,h<'1;C douhlc-r/Il<l 'torlOise' mOli(~ arc .ltcrnath·d,kno"n as moh. i e. 'Iilard~' "lIi"b [11,,\ do indeed mUt;h more ,,-,;emble b) yinlle oflheir CT<J<>l;cd limb.and d"ngall"d bod)" shape. One of t~ IInsohcd m~"criC!l of the lilerJlur<: on Polynesian IJlIOOiog i.

180 SlY/I! lind Cullufl' 181

FIG. 8.10/2_ ,UIlIIl hllllla in male"uoos rec(lHltd by Steincn.5011fa: Von den Sleinen, i, ii.(\"1lrillu~)

ia vau

~~

Fin. 8.10/4' From mala/Willi' 10 rhe Vallrock modf b} (ht addition of spimls,SOllfa: Von dcn Steincn. i. I(M),

inus. 125

F[(i. S.IOI3. The mill" /JlNlID t,mooindic:ned lends e\"identl~ IO"":lrds Ihenil" motif SOllrrr. \·00 den Slc;ncn,i. 1

FIG. 8.10/'. Silling (Iua hypostasizcd II) Ilwta Iwata.

Tladora's leg f;\ttoos: (a) lIIaw hoall'; (h) double Imkr;(c) matll hoalll; (d) l'alliupa; (c) Elua; (f) IIWla Iwalll;

(g) double kak(; (h) poriri. SOUfcr: Von den Srcincll,i. 132, illus. i7

8.10. From the Erua to the Faa Motif (Mata Hoara)

The next transformation of the /!lila motif lO be considered i .. perhaps the mostfundamental one of all to the integrity of the i\larqucsan style. This is thetransformation of the sitting tlila into the 'e~e' or 'fat.-c' motif, mata IlOa/a. Theface motif is produced by hierarchization or hypostasization of the t//IO motif,with which it is morphologically !qlC3king mtJTe or Il;Ss identical. The derivationi<; most perspicuously seen in a \'ersion of mata hoala ('shining eyes') featuringon the leg tattoo of 'Frau RadoI"'J.' (Fig. 8.10/1). Badora's mata hoMa featuresa particulad)' explicit eltlfl (identical to pollll = 'r' on Fig. 8.7/1) forming the'nose' element above which arc to be scen the large half-{)\'als forming theprominenT eyes of this face design. Placed thus in conjunction, it becomes par­ticularly apparent that the maftl hOil/a design is identical to the (lila (pO/III)design, of which it is a scaled-up version. 'fhe outstretched arms of the If/IIII motifbecome transformed into the deep curve under the eyes, rhe 'hands' of the etl/abecome rhe irises, and the head of the ttlla becomes the bridge of the nosc. Moreusually, the 'nose' of lIIala hoata is a transformed version ofpohu which is notso legible. Fig. 8.10/2 shows some more standard forms of mata Iwaw frommale tattoos recorded by Steinen. To produce this t) pe of nose, the 'arms' ofsmall ~/ua are rcflected horizontally so that lhe) form opposed half-{)\'als, andthe head is e1iminatcd. Thus a shape morc reminiscc", of nostrils is produced.Howe\'er, mata hoata as a whole remains a legible derivatiyc of the basic duamotif. Further rcflccred hal(-o\'als may be added 10 the cUlTcd cye margins to

add extra potency to lhe design. Whcn mala ltonta are inscribed in restrictedspaccs, the motifbccomcs increasingly synonymous wilh el/la (Fig. 8.10/3)·

ared from (fl/a set end to end have 'shells', these arc represented by the secondand third sets of 'legs' of the resulting eight-legged creature.

In 18o-ll.angsdorff published an engraving ofa youth with a re1ati\'cly 'real­istic' tortoise t3UCMX:d on his elbow; in this \"ersion of the tortoise motif, thecarapace is drnwn as an oval, ami the tortoise itself as an CIlia contained withinthis O\"al (Fig. 8.917). If we compare Ihis with (he form Jua (tortoise) asrecorded by Steinen in the 1&)05 the carapace has disappeared, and we ha\'conly a row of ellla motifs. But perhaps not entirely, because the way in whichthe Icgl' of kca are drawn (large quarter-ovals) suggests that the 'cllrapacc' oflhc origin,lllOrtoisc motif has heen divided inLO quarters and rediHl'iburcd as'legs' for the eflla-in filct, adj:u:enr fllla share a half-shell between them.

why rh~ famOIlS Ml00 an ofl~ X~l\" 7.e-~bod J\buri should h~,c been known as moho Tht prtscm'tof ;"01:0 as an ahtro'li>c dcsign~(ion for I:~o (l-"k~) in ;\lor'luCSJn tal{l)U is ",,",inh sugg:C$li,~, l.i7.ardsw...-." lik~ ",rluises, sacred ""'inp, 3nd Ihe~ m,L<' numerous al'pe'~"'Jncc.< in \Iarquo:sa" UI. The WJ} inwhich such dissimibr repliles 'IS rort"i"""" ood linrds Ita,-c bcconlC ,iSU311~ synonymous in .\brquesan3n I"'-nl~ ","O..'''ls lheir similar sJrnoolic 3SSl.W,.-i.llions.

182 Style (md Cu{llIr( IS3

If'l"l"l' 'f 't"!' 'f"!, 't'

FIG. 8_11/.... 'H:lnd r~{T' design J: o::tch e~·e em:!.Iso br Ic::I.<I :a:;:l fxe. one: single tlua, or:a:; one(Iuu conl:!.in;ng :lI101her. Sauret: Von <1':0Sicinen. i. In. iIlus. 13616

JlOlltalipu tJOlnsformation is sho\\ n in Fig. 8.1 lit. This iplI motif, bisect cd hor­izontally. immediaTely transforms inTO mala 1IO(I(a. Usually, ipu arc surroundedhya mOlif, p"/llIa, which is a reduplicated derivative of the 'nose' element inmala Iwllla (cf. Fig. 8.1O(2). Although the motif-name ipII does not refer toeyes, iplI arc frequently placed so as to m.lkc this identity palpable (e.g. theunique tapa mask shO\m in Fig. 8.t 1/2).

Next, I turn to the family ()f'face' designs other than lI/fIta Iwata which fea­ture circular eyes. A common site for the placement of this C".llegor~ of designswas on the hands. the b;lcks of \\ hich were protectcd by 'faces' composed of avariety of elements. Fig. 8.11/3 sho\\s such a 'hand f:.lce'. It \\ ill be seen ThaT

J.)G. 8.1 I/J. '11~ndr:lce'design I.Sau,(r. Von den Sieincn. i. 177.illus. IJ6/5FIG. S.II/J.ltft. T:lllQ

mas}; ..-ith IpU ·c~cs'.

SOllrer; Yon denStcinen, i. 17J.iIllis. 131

FIG. 8.11/1 ,jil' leji.From double fIIulu

11mI/O 10 ipu. S"lIrer;Von den Sleincn. i.129. illlls. H

)

Adding spirals to the ends I)f the efua's 'arms' and on each side of the headproduces a further motif, similar in its proportions to mula IWUIti, called vall

(rocks). This motif rCL-alls tile ocean swdls breaking over coastal rocks andwas panicularly placed 0\ er the buttocks. Vall is thus an additional dcrh"ati,·cof ~JualntafQ honla, though ttl lend the motif grL':ltcr ,"erisimilitude., it was com­mon to disrupt the symmetry of these spirals in various ways, producing theramil) ofahcrnariw forms 5hOl\ll in Fig. 8.1014.

8. T r. From Mara Hoata to Ipu: AdJitiolllll Fare A10tifi

Doubling-up the 'eye' half-ovals of mata hOI//a produces the next importantmotif, ipll. Ipll means 'bowl' t)]' calabash, and it was usual to place f()\VS ofrhesealong rhe underside of the ann, where they would be rolnged in polirs. The lII(J(a

J\.) i

184 Style <Iud CII!JIlrt' St)'fe alld C/lflllrt" 185

here the eyes :lfC composed of two opposed half-elull figures, placed over a'nose' element identical in form to that seen in mala Iwata. The 'mouth' of thehand-face is composed of:l transformed version of the deep 'eye' cun'cs ofmata hoala. An alternative 'D'-shaped rathel" than circular eye form is seen inthe illustration in Fig. 8. [ I f4. Here, the eye is formed by a single ellla, t!lOughthe right-hand one of these has a small ~ldditional etua-figure encapsulatedwithin it, and both verge towarus becoming miniature versions of !/lata hoata.In other words, this single motif can be read simultaneously in three quitedistinct ways: (i) as a single eye of the hand-face; (ii) as a face by itself mala

lioa/a; and (iii) as all elua, or indeed, as two elua, one containing the other.This overflow of possible readings attaching to a single motif-surrounded bydozens of others, equally perplexing-perfectly instantiates the cerebral wit ofthe Marquesan style. What was the point of such displays of graphic inven­tiveness? That question must be set aside until the entire range of stylisticch:lracteristies of Marquesan art has been explored in more detail.

A common face motif, seen especially on the knees in female tattoo, iscalled kaufllpa or poriri ( = 'coiled shellfish' according to \Villowdean Handy).Steinen himself designates this face motif as 'kake-type' since as Fig. 8.lT/sshows, it can be regarded as a I.·ake (podoid) in circular form, though the basic

FIG. !(II!S. 'Coiled ~hdlfish' poririas a face motif. Souru: Von <.len Steinen, i. 179, i1llls. 137

kake shapc has been considerably modified to make the 'mouth' and 'eyes' ofporiri. None the less, Steinen's suggested derivation seems well warranted.

Lastly, we may note that filce motifs can be hierarchically subsumed into largermotifs, and may also subsume smaller ones. Thlls the 'eyes' of Fig. 8.11/+are also faces. The same idea is exploited on a larger scale in Fig. 8.1 1/6, anengraving from Lmgsdorff showing the (unfinished) h:lck-design of a youngwarrior. Here the two large circular designs over the scapular area arc 'eyes'(the eye design is similar to the one seen on the tapa mask, Fig, 8.11/2). The'nose/mouth' of this back-bee is composed of the smaller face in the small ofthe hack immediatelv below. This recalls the point made earlier wilh referenceto Fig. 8.9/+ showing the 'girdle' of a chiefess, where the eye elements arc kake

enclosing oval eyes made up (rarher randomly) of elUi' forms. Here the nose isthe eflla in the centre of the design.

FIG. 8. I r/6. r.an",,.,lorff's 1813 en"'Ta,,ing of young "arrior \I ith eye and hce dcsigns. .'iouy,,":LangsdodT 1 R'3-q

186 Slyle anrl Cullure Slylr lillii CIII/ure 187

8.12. Some Additional Mot(I~Trallsform({lioll

The few remilining Olmmon motifs can be dealt with briefly. A very common'tiller' motif is papua (see Fig. 8. T Tf4), which features small half-ovals in a vari­cty of configurations, andfor the 'nose' element from mata Iwata, which, as wesaw, was a small transformed etlla motif. A motif for which I can find no satis­factory derivation is pallilo which is also called pORaRa (a package or bundle,which it does indeed recall). This motif is formed from opposed half-ovals invarious dispositions. In practice pahilo just delimits a space, whieh is usuallyfilled with derivatives of hope vd/ine or mata Iwata. Another common motif isthe roseLLe (.leo '0) which is made up of row motifs (such as forms of poilu orjima1!a) arranged ill a circle (Fig. S.I2/I).

Finally, I shall only mention some of the smaller motifs which Steinen dis­cussed under the rubric of 'plaited' or woven motifs. Some of these consist of

FIG. 8.1211. Rosette motifs.Sourrr: Von dcn Stcinen,i. qX, illus, 93

FIG, 8,1212, ft/i. I-1icrarchi~ation: niho 'tcclh'motif;;

FIG, X.12/3, he/mr, A"i'lliu band eomainingaltcrnating rc"crsct.l (IIUI motif I in Fig. 8.7{1.S'IlIrre: Von den Slcinen, ii. 19,t, illll~, ")2

various arrangements of triangles, identified as 'teeth' or boniLO spines (teethcan be simply transformed into spines). A common use of teeth motifs (lliho)is to 'armour' the triang'ular patches of solid black tattoo on the inner thighs, inmale tattoo (Fig. S.I2/2). These triangles ornamented with triangles providea particularly transparent instance of hierarchization. Other row-motifs arethe small elua figures, poilu and Jimaua, that have already been mentioned ascomponents of larger motifs. Their main use, however, is as fillers, arrangedin single or multiple rows (they arc shown on row 3 of the table of etua motifs(Fig. 8.7fT). An interesting variation in such small etua motifs is achieved byjoining them up, either in the same orientation, or in alternation rotated by 90degrees. (Fig. R.1213 shows lhis alternating band pattern in the motif aniatiu,from the table of etlla '1'.) Finally, the most extreme simplification of the etUIimotif is the chequerboard pattern. Just how this pattern can be derived fromthe 'standing' etl/a motif is made very apparent in Steinen's table of eilla mOlifs(row 2, a-I, showing variants: the symbolic and political significance of thechequerboard pattern is discussed in Gell 1993).

8.13. Figure-Ground Reversal (Tortoiseshell Diadems)

This is as much as I wish to say about tattoo motifs. The next category ofart to be discussed is l'vlarquesan work in tortoiseshell, which t{lrms a bridgebetween the two-dimensional art of tattoo, and carving, in relief and in Iheround. The most delicate shell work is entirely two-dimensional, and consistsof'diadems' (Fig. 8. T3/ T) in which a findy incised disc of torroiseshell is shownoff against a slightly larger disc of pearl-shell, forming a most elegant orna­ment for the forehead. As Fig. 8.1]/1 shows, there is a great deal of continuityberween Ihe style of Ihis pierced shell work and tattooing. The disc I illustrate(Salem 1807) reminds one,;u first glance, of mata 1/011/11. But exactly how arethe faces one sees arranged, and what are their separate components? Aroundthe central aperture, there arc four large 'eyes'~enough eyes for 111'0 EKes,normally speaking. Given the cifel! Jar layout of the design, however, these foureyes belong, not to TWO, but to four faces, each of which shares its eyes WiTh

188 Style lim! CII!llIfl' Slyle (wd Cut/un 189

FIG.8.IJ/l. From cyes 10lizard: Steint'n's cvolutionaryderivation from end-plalc discs.Sourer: Von den Sltinen, ii .'73. illus. If)!

•

*

*

*

m

ErUA

p

*

*

,fIB

,

";';;'~':;(\.:., ....~, :. -: .....~~(. '. ..,

l-)G.8.13/1. Fil6Un:: ground~nrsal and hieran:;hiZ:llion:tortoiseshell indsed diS(:SO/lrer: Von den Stdllen.ii. 168, i1Ius. 156 (i"J

its neighhour on either side. The large eyes, read from The centre of the dis!;,arc composed of clements which look anthropomorphic-a figure with r.lisedarms, or half:1n tWa.

Bet ween each p,lit of eyes, there is a complex form which constituTes Ihe restof the face. R<''ading each f.,ce from the rim of the disc inwards, one can m'lkeQUI nostrils and a mouth. Ilowcver, if one reads these mouth/nostril parts fromthe centre of the disc Qmwards, a quite differenr set of faces emerb"C. Wh:u werethe 'nostrils' of the big faces s(:cn fmm the rim of the disc, hecome the e)ts ofsm.lller, skull-like faces which are seen from the centre of (he disc (Fig. 8.13/'),Nor is this all, for these smaller skull-faces set between the large eyes (whichone sees in white-on-black) are themselves created from standing etua figures(which one sees in black-on-white). This is a stunning example of figure­ground reversal, much more interesting than the hackneyed faceleantllcstickone cited in all the psychology textbooks. As I will show later, such \'irtuosityis far from the arbitr:ary exercise it might seem, and relates directly to certainfundamemal attributes of the Marqllesan cllllural system-bllt at the sametime, 2S I noted earlicr these attributes only bCl..'"()mc thcmatic aftcr onc hasuIr.:en the trouble to dissect the style and morphology of the material culture,not before.

Diadems of this kind were made in three pieces, the central disc plus twoend-plates of the kind shown in Fig. 8.13/2 (some diadems had three discs ~Uld

two end-plates). The end-plates make an interesring study, in that one can seehere a transition between a 'two large eyes' type (whitc on black) transformingitself into something apparently quite diflerent-the image of a lizardjt'111fI (inblack-an-white). There is an obvious relationship betwccn this figure groundreversal :\11(1 the one discussed in the preceding paragraph. The evident 1fan­sitioll between 'eves' and 'lizards' was grist to Steinen's evolutionary mill, andone can readily s-ee why (Fig. 8.13/2). At lhe same time, onc mUSI rCl.:allthatthcsc end-plates were nOI designed to be seen grouped together for comparison,but separalely :lnU individually, so thaI the transitions that we can sec would

not necessarily have been ,\pparent in the same way to Marquesans. Onlythose visual ambiguities which arc co-present within a single eml-piecc are ofunquestionable stylistic s:tliency. 'rhere arc many such ambiguities, howevcr(for example, '0', 'h', where eyes ami "Iua/lizards clearly {;()cxist).

Perhaps the most' signifIc;lnt :lspect of these cnd-plates, in terms ofMarquesan art as:1 whole, is the light they shed on thc transformation of lhefrontal (11111 figure into the 'profile' (IIW. via the 'lizard'. Tn tauooing, 'profile'(Ilia arc not to be found; lhey only occLir in 'miniature' forms of plastic arr,such as thesc cnd-plates, and p;u,tieularly in carved i\'ory earplugs, which 1 willdescribe later. The gcnesis of the profile-ellUl form is easily seen in cnd-plate

8.14. From Two 10 Thru Dime11sions

'p', in which the creature at the ba!':C could equally be a lizard-type elllll like theonc in 'r' or two e/llll with ben! knces seen in profile. I shalt ha\'e more to sayabout back-to-back divinities later, since they arc the dominant motif in allforms of Marquesan plastic art in the round. In the end-plates, though, theyare commonly subjected to a transformation whereby one of the back-to-baekprofile ttuo is reflected, so that both now face the same w;I,y (Fig. 8.13/2, 'i';both \'ersions of the motif occur simuhanL,,(lWily in '01').

Now it is time to turn to Marquesan plastic art in more than two dimensions.I think it would be fair to say that no Marquesan art is fully three-dimensionalin thc sense of fillly occupying three-dimensional space. Marqucsan sculpture isconceived :IS a 'skin' around a corc (a cylinder or cubc----(lr a human body inthe case of tattooing). The core may be three-dimensional but the can'ed designoften seems to be wn(,;eived in the flat, in low relief, as applique decorationwhich is \\Tapped' around this core. This even applies to large free-standingJigure sculpture, especially the heads. This raises an imcresting point aboutexactly how plastic art in three dimensions is related to graphic art in twodimensions, of the kind we have been considering up to now. The Marquesansconceptualized the transition between two-dimensional anti three-dimensionalart in a way quite contmry to our own habitual way of thinking, as I shall nowattempt to explain.

Suppose I prescnt you with the image of a person, drnwn in inl.:., on a sheetof paper, and I pose the question: 'what is the back view of this image?' In asense, this is a non-question, bCl."3use drawings are graphs in two dimensions(up-down and left-right) so they do not possess 'backs' or 'fronts'. We, how­e\'Cf, naturally interpret Ihis: question as equi\'alem m: 'What docs the backof the person you see in the drJwing look likc:'--or perhaps, if we choose tobe literal-minded-'What does the back of the piece of paper on which thisperson is drawn look likd' Since the question seems: a silly one in it,. liternlinterpretation, it would be n:llur.:al for us to assume lhat lhe question is oneabout what people look like when SL"Cn from the back, and we might answerit by producing a drawing like this (Fig. 8.q{r), which is the same person inthe same pose seen from the rear. \\ie do this bec:lUse we 'see' drawings as 'flat'projections of the three-dimensional objects thm dnl\vings represent, especi­ally when these arc dr:lwings of (1mili:lr objects, such .IS human beings. Thisvisual presupposition is the outcome of centuries of conditioning to natural­ism in artistic representation; :IllY drawing is a 'drawing of' something whichexists separately from them (the 'subject' of the drawing) and thc subject of adrawing is typically a threc-dimensional object with ;1 frOnl and a back as wellas a lefL and a right and a top and a bottom.

These presuppositions do not hold for Marquesan art, which was not 'rep­resentational' in our sense at all, without being 'abstract' either. Marquesan artwas a ritual art whose purpose was to render the person powerful and invul­nerable. I will return to this subject in the conclusion of this chapter. For thepresent, it is sufficient to say that the tanooing of, for example, an elua motifon the body was not a matter of representing an ttl/a which existed (as a three­dimensional solid object) somewhere else. The tattooed elua was protective ofthe person because it /PtlS an etl/tI, right there on the body, not because it'looked like' an eJlw somewhere else. The graphic act was a ritual performancewhich brought into being a protective spirit through the utterance of a 'legit­imate' (stylistj(.1tlly coherent) graphie gL'Sture. One cannot speak ofgraphic ges­tures of this kind as representational in our sensc; they are constitutive. Thisis the c..-ssencc of Marquesan idolatry (and of all idolatries), which informedtheir entire approach to graphic and plastic art.

Let us now reconsidcr the question about "hat the back view ofa drawinglooks like, bearing in mind that a Marquesan graphic gesture dOl."'S not repres­ent something, but constitutes something. An ttua motif consists, in ilself, ofa bundle of lines. Ifwe SlOp thinking ofthcse lines as 'representing' somelhing(else) and just think of these lines as lines, then the answer to our initial questionis plain; the 'back ,'iew' ofa drawing of a person, is that drawing seen from thepoint of \·iew of the paper, that is, the mirror-rc'·erscd "iew of the same drawing.

This proposition may be applied to the gencsis of three-dimensional plasticart in the Marquesan style; whereas Western graphic art is the representa­tion of 'tl;ltlened' three-dimensional objects in two dimensions, l\'larquesanthree-dimensional art is thc projection of two-dimensional 'flat' motifs inthree dimensions via the mirror principle, thaI is to say, the back Of:l rhrce­dimensional image is the mirror reflection of rhe front. Marquesan rnoti[~,

which arc constituti\'e graphic acts rathcl" than aCls of represenration, ha\'eno 'backs'; all the ;Ivailable \·iews arc canonical, bec:lUse the principle of ritualefficacy depends on lhe stylistic legitimac) :md 'correctness' of the gesturc.

191

FK•. 8.• ",,1. From Ihrttdim~nsiollS to 1\0'0 inW~ttm :l.n: the b:l.d: \'ie..·of Iht im:l.gt ~ lilt b:l.tk,·it\\' or Inc, prOU){lJX

Style alii' Cull lire

(~@ ~J]"-{ VT\ c('t-'-.-

oij Jth

I

Stylt and Cut/lireIIJO

192 Slylr 1IR11 Cuflur~ Slyl,' IJnd Culillft 193

8. I 5- BO(J:!essl1ess 011t1 Split Represmtalioll: Sltell Crowns

FIG. X.15!1. From two dimensions to thrce in ,\brqllcsan art: prOI'iding the image "'ilh a hack.S",m(: Von dcn Stcinen, ii. Ij6, i1Ius. 1M (.1\) and SI",in",n ii. In. illus. 1('7 (D)

One C'Jnnot see God from the back, because if he is not watl,hing us, he is nmGod. In the s,1me way, Marqucs:m art cannot permit motifs to be witnessed, soto speak, al a disadvamage (just as it was sacrilege, in the ~'larquesas as in mostPolynesian societies, LO approoch, or pass by, Ihe back of a chief).

example of a plate consisting of jusr one t:tlltl filling The whole Spacc. I argucthat this is to be auributed to the need to ensure that these tllla images, whichare fct:c-sranding, should at least conceptually be prm-ided with backs mir­roring their fronts. These images arc, as it \\ere, folded outwards so thattheir backs arc nm\ in front. Thus, if we were to imagine that Fig. 8.15/1, 'A'were made of Oexible material, we could turn the IwO outer 'profile' 1'/1/0

faces around the back of the central one, so that they met in the middle, wherether would join up to make a back elua mirroring the front one. In otherwords. thc.coe 'profiles' are half-faces which I,;an be put IOgethcr again. A sim­pler solution 10 the same problem of providing a back for a free-standing ellia

is seen in the alternative layouT '0'. Hcrc there arc two mirror-reflectedfl/la which, on the same basis as before, could simply be folded down themiddle to make a back elua and a front eluli. All the tortoiseshell crown platesare \·ariations on one or othcr of these solutions to providing two-dimensionalrelief carvings with (protected) 'backs'. Howe\-cr, Marqllesan concern withproviding ell/a images \Iith symmetrical backs did not end there. Some plates(ii. 182, illllS. 174) ha\-e, in fact, carved backs identical to their fronral de­sign, even though, in use, these car....ed backs to the plates would ha\·c beenmostly, if not entirely, im·isible. Laboriously adding dccoTation to \'isuallyinaccessiblc surfaces sl,:ems a pen ersc waste ofeffort; bUT it was necc.cosary nonethe less if these crowns were to provide rilual prol.ceTion from all dircctions,as intended.

As will be Sl-'Cn, elulJ faces and profiles sprout everywhere from thesc plates;often the bodies/limbs of the main etufl become suhsidiary clua, <lnd so on.Steinen makes:l dctailed study of the yariations in this regard (ii. 174-82). Asimilar analysis to the previolls onc of tattoo motifs might be undertaken, butI shall refrain from providing onc. The last point I wish to raise with reg'J.rd tothese lortoiseshell plates is the Lloasian/LCvi-Srraussian one of 'split repres­entation', which narurally Stcinen does not discuss (he citcs no non-Polynesiancomparath-c examples). It will be apparcnt That both typc.<; of plates, the 1/2 + I

+ 1/2 layout and the 1 + 1layout, are instances of'split representatiun' of a kind,cognatc to, though by no means the same as, the farnow. examples from theart of the north-\\ est l"Oastal tribes of America (Fig. 8.15/2). J\ larqucsan splitrepresentation is primarily in the plane bisecting thc body from lcfl to right(as in D) rather than from back to front as in the American/ancicnt Chineseexamples discussed by Levi-Strauss (1963: ch. Xlii). Occasionally therc is backto front splitting as well (as in Fig. 8.1 SIt A, though thc 'halves' ofthc split rcp­resentation Elec away, rather than Towards, one anothcr), and stillmorc rarely,there arc instances of split represent:ltion where the 'halves' face one anotheras in the I.cvi-Snauss/Boas-type split. repreSelllalion (e.g. see ii. 179, illus. 169a). But the precise geometry of split representation is perhaps less imponant[han the inrcrpret:ttion onc giyes [() it. Lhi-Strauss idcntified split reprcscma-

oA

Toe mOSI perspicuous instance of backlessness in i\hrqucsan art is providedby the tortoiseshell 'crowns' made of a series of cun·cd, rectangular platcs ofshell caned with etua images of great richness and intricacy (Fig. 8.15/1).These l-TOWns are the 'exl-"Cption which proves the rule' in the sense that theyare low-relief, two-dimensional carvings Tather than three-dimensional can·­ings. The plates are, for the most part (but see below) only decorated on oneside, the side facing out. Yct in fact, they arc tct.'aled, conceptually as if the)'wcre works in three dimensions, Marquesan fashion, that is, with backsmirroring their fronts. This may be seen from the phlle illustrated (ii. 176,illus. 166 'A '). This is a typical example in which the main motif is an 1!l/1tI, seenfrom the front, full length. But the main figure is tllllked by twO fun her ell/{/.

figlll'cs seen in profile. The Olher Slandard layout for such plates is seen in'D'; herc there are just two tllla, I,:ach one of which is a mirror rellection ofthe other (see the arrangemenl of the arms). J\'leanwhih::, there is nOI :1 single

194 St)'/1' and Cul/ure SO'II' 11m! CU(/lire 195

tion in a rather forced way, I think, in Maori art, essenti:\l1y oc'(.'lluse Maori rep­rl.."'>Cntations of the f.,ee, C'Jr\"cd as reliefs, tend fO open the face our so that itbegins to resemble two opposed profiles rather than one single face viewedfrontally. This characteristic is equallv if not more e\'ident in ;\ larqucsan rep­resentations of the face, but the doubiing of images, in the tortoiseshell platesI ha\ e just discussed, is perhaps a more perspicuous example of the 'splitting'principle which is rc-J.lIy what is at is.o;ue in split represcnration.

FIG. 8.15/:. FromLhi-Stnuss 1963: ~..IilreprCSCTltatiun and theJanus-faced image innonh-wcsi !\mcril.m tribalarlo SUlmt: Lh-i-Strauss1963: fip. 17 and 18

Uvi-Strauss's interpretation of split representation is most thought­provoking, however, in the Polynesian context. He maintains that splirrepresentation occurs when social factors militate against the dissociarion oftwo-dimensional and thn..'(.....<Jimensional images. 'In the end, our pmblem maybe fonnulated as follows: under what circumstances arc the plastic and graphicL-omponents [of an art systeml necessarily correlated?' he asks, and responds tohis own question by asserting that the relationship between the plastic and rhcgraphic 'has to be functional when the plasric component consisted of the faceor human body and the gf;lphic component of thc facial or corporCllI decoT"d­tion (painting or t:mooing) which is applied to them' (1963: 26t),

A face or body in its two-dimensional aspect as the eall\"3S for'decof;ltion'(i.e. magiC'dl graphics) cannot be dissociated from the three-dimensional bodyto which 'decomlion' is applied. Hence, argues l.cvi-SlnlUSS, split repres­entation is associated, not just with masking cultures, but particularly withthose masking cultures in which mask and person ,Ife indissociably linked­as is most particularly the case with body-painting and especially tartoo~n~,which is quite irremovable, Recalling .Mauss's discussion (1902) of the onglllof the concept of rhe 'person' in the social mask, person,t, one might say thatwhere persona and person unite, representations arc split because ~he t.hree­dimensional person and the two-dimensional persona (,mnot be dISsocl,lted.Lhi-Strauss says lhal societies which are both \"Cry hierarchiL-a1 and very com-

petitive arc ones in which these conditions arc fulfilled. Such societie.'i rypicallvcompete o\'el' genealogical credentials linking men with gods, as was lhe e~among rhe tribes of the north-west coasr of f\merica, among the hierarchicalsocieties of South America, the archaic societies of the Far East (producers of~e Shang. bronzes, whose affinities with Polynesian art have been noted b\'E. ?ombnch .1~8.r 262(0), among the J\hori, and, equally, the ~'1arquesan~.ThIS compcllt1\'e struggle to assert genealogical status is the sociologicalrationalc behind split representation.

. I would ~O[ ~'llre to ~rgue for thc un~v:rsal validity of Levi-Strauss's sweep­mg gencf;lhzatlOn; e\'ldcntly, eompettllOn ovcr genealogical credentials canbe engaged in without this finding expression in the particular modalitv ofmasking, split representation, or anything of the kind. However, What I~evi­

Strauss says about the sociological characteristics of'split-representation soci­eties' is quintessentially true of the l'vlarquesas, and there is indeed somethinguncanny about the precision with which he rhetorically anticipates Marquesanethnographic data of which it is most unlikely that he had foreknowledge,Summarizing his point of view, he remarks:

splir reprcsemation expresses rhe strict conlilrmity of the actor to his [slXiall roleand ... to myths, ritual, and pedigrees. This eunformity is so rigorous thai, in orderfor the individual to be dissociated /Tom his social role, he must he torn asunder.(rl}63: 26..)

I say this is uncanny, because, as I have described elsewhere (1993: 2t3 IT.), ifwas acmall)' the case in the _\larquesas, that dead chiefs were 'torn asunder',and with this precise purpose in mind, that they should ha\·e remo\·ed fromthem the social identity conferred on them hy tlleir tal/ooing. The tattooed skinof chief'i was removed because, in .;\ larquesan belief, dcifiGltion and a pleasantafterlifc were denied to the rattoocd because of the association bet\\ccn tanDO­ing and the mortal condition. After death, the skin reinforced by tattooing-sonecessary to ward oITthe physical and spiritual dangers ofa worldly and chieflylife-had to be left behind; social skin and immortal soul parted compan~ andthe chief rejoined the assembly of the c1ear-skinm:cl gods.

I hope that the eOIl\"CTgcnce between Levi-Strauss's argument on the subjectof split representation and the one advanced earlier concerning thc pen:lsivetendency towards providing three-dimensional images with 'backs' reflecting,mirror fashion, their frontal aspect, will be appreciated. The prohlem of splitrepresentation and the problem of Janus-faced images-the latter bcing aparticularly prominent characteristic of Polynesian plastic art arc of :In essen­tially identic:ll nature. Janus-ElCed images are [0 plastic art what split repres­entation is to graphic arl, that is, the means of securing that I/O pari olall imagedeparls from the '((/llollim/' (riluul6' po/elll) .';em. In BoasfLevi-Strauss splitrepresentation, if an animal is seen from the front. then the whole of the

196 Slyk IJlld GUltllft Styh IJlld CI/I/li.ft 197

animal is seen from the front, including tho~e parts of it (the flanks) whichare in f.1ct invisible when a 're-J.I' animal is seen from such a vantage-point. InMarquL'San ('Polynesian') split repre.;entation, the strategy is slightly different;the image of the front is reduplicated and become... the image of the back aswcll. Where split representation (a graphic mo<.le) pro\'ides a canonical "icw ofa lhree-d.imensional objecr in a t\\o-dimensional space, the making of Janus­f:aced images allows for the expansion of essentially tw<KIimensionai graphicimages into thrL'C dimensional pl:astic space without contra\'ening the stipula­tion that 'only canonic images arc permissible'. Hence: Janus face: split repres­entation:: graphic art: plastic art:: three dimensions: two dimensions, etc.

8.16. Janlls-Ptlced Images: Scale Transfiml1aJio1lS

However, in the course of the preceding digression inro the problem of splitrepresentation, I have perhaps run a linle ahead of myself, in that I have nOlyct introduccd the full panoply of Janus-faced imagcs with which ntarquesanplastic art in three dimensiolls is STrikingly replete. To the description of thistype of image I will thcrcforc tum without delay.

It would not be true to say that every singlc one of the anthropomorphicsculptural images produced by the !\'larquesans were in what Steinen describesas the 'doppeltiki' (double tiki) form, but the great majority of them arc. Theones lhat arc not include a small number ofstone images, with uncarvcd backs,some wooden 'POSt figures', and a larger category of small wood and boneimages which, though norrrueJanus-type figurines, arc carnd with subsidiary(protective) baek-t'II/O which 1 will describe in more de(:I.il below.

T.arge stone images comparable to the famous Easter Island statues werecan'ed by the Marquesans, though the) nevcr attained the num~rs, or thesize, or the magnificent workmanship of Ea...ter Island art. I would attributetheir lack of hack-to-from symmetry to the fact that they were probably alloriginally built into the stone terraces of chiefly mara~ (ceremonial platforms)rather than being free-standing (Fig. 8.16/1). The most interesting of thesestonc images is the one known as :\1akii-taua-pepe (Fig. 8.16/2), which rep­resems a female divinity gi\'ing birth. There are other images of birth inMarquesan art (in the form of miniature carved ivory car-ornaments) but noneon this scale. I will discuss Marqucsan ideas of parturition when I COme \0 dealwirh rhese ear~ornaments. for rJle present, it is sufficient to note that despiteirs unique form, this large statue is conceptually a 'double tiki' as wcll, in thatdming parturition (a momenr of great danger ,llld sacredness) J'vlarques:mmothers physically ;lssumcd (he 'doubled' form, otherwise associated with divinebeings, and children, produced in this way, were themselves very sacred (Gel!'993: 185). In this c,lrving, the 'divine child' is represented, as Fig. 8.[6/2shows, by a standard ell/II mlltifon (he undersidc of the distended belly.

Fla. 8.16{1, kji. Stone rcrr:ocr (Ifccremoni:l.l pl:nform. Sf1NfCr. ,'un denSleincn, ii. 'j"~ ilIus. 55

FIG, 11.[('/2. iHakii-laua-pepr: !COl'lledivinity giving birlh. Sou,ce: Von denSleinen, ii. 81, illus. 61. Sou,a: Vonden Slcinen, ii. IZO, illus. 95

198 .·:;~)'/r ulld CII/lllrt Sty!, I1l1d GlIllurr 199

FIG. R.I6/3. Wooden JlO5IS cu"'w inl ....o-dimC1ls>On~1 form. Slmrcr: \'on d~n

Sicinen. ii. 100, iIlus. H

Ffllm the point of \'Ie\\ of 5ty1istlcs-whieh is our primar} concern-thepoim to nore about all these SlOne images is their block-like form, and the wal'in which the fL'3lUrCS, especially those uf the face, are not modelled in thre~dimensions, hut arc incised on the block in low relief. .\1ore or less the sameconsiderations apply to the wooden anthropomorphic images (Fig. 8. I 613).These arc not free-standing images but wmponents of houses or other struc­HIres, such as c:lrved posts, supponing beams, which, as with the stone images,explains their uncaned backs.

The disproportionately large eyes/mouths in all medium- to large-scale .\ lar­qucsan sculpture, and the absence of noses, suggests to me thai the stylisticconventions for work at Ihis scale werc deri\ed from work at a much smallerscak·. On:l tiny bone can'ing, fearures of these proportions seem quite natural,the result of the need to delineate eyes and mouths visibly, which would beimpossible in miniature work were the natural proportions of 'real' eyes ormouths observed. Nose,;, on the other h:lnd, are redundant at this scale, sincethe tri:lngular SP:lCC between the enlarged eyes and moulh implies the naS:lI tl'i­angle \\ilh no need for morc Ihan a summary tre-.lImem of the nostrils, and theabscnL'C of a prOlruding nose coincides with the essentially two-dimensionalnature of ~larqucsan carving, which is incised in 10\\ relief rather than beingrruly 'carved'. To my mind, all the Illore imeresling Marque&1n plasric art,from the stylistic point of view, is al the small or even miniaTUre sc'lle.Marques:m arL, as I suggested earlier, has to be understood as a Technique forenhancing the person b~ the addition of spiritually potent appendages orsupernumeraries, ;111 essentially in the form of rilla. The larger al'l forms wereappendages ofstructures, such as ritual platforms or houses; as such, they wereen1:lrgcd equinlents of 'personal' appendages, such as fans, amulets, headornamems, ear-ornaments, sporting gear (stilts), and \\eapons (duhs}--and ofcourse, lallooing. It was in rhe domain of these 'person-cnhaneing' art formslhat the basic forms of Marquesan art were generated, so it is 10 Ihis area thatwe must devote most of our attention.

Once we eTHer this domain, the presence of the 'double tiki' form becomeso\-erwhdming, beL1lUsc, as I ha\'e noted above, the requiremenls of magicalefficacy in plastic art forms (i.e. three-dimensional ones) demands the simuha­neous presemation of the canonical (efficacious) image in all spalial dimen­sions. A suitable point of dcparrure is pro\'idcd by the small bone tiki whichwere used as hair ornaments. These are cylindrical in form (Fig. 8. I 61-t). Theycame in two patterns, either fully symmerrical or, more commonly, \\·ith a sub­sidiary back-figure. which in turn was eiTher a sm:\11 elml mOlif, or a squal'ecngra \'cd with .t IwJ..ellkreuz. 'I'he hakenkn!/lz motif is ;1 \'"rianr of the flila (Fig.8.1615) \'ia hope whil/e (see abO"e).

200 Slyle Ilnrl CultureS~}'k mId Cllflllre 201

more unnatural head/body/legs ratios: (2) A = H = C, (3) A > B:::: C, until theratios found in these bone liki arc attained, at the opposite pole from 'natural_ism' (4) A > B > C

In the plastic arts, the head as an isolated entity is very rarely found; every headmust have its body; but it may be given very meagre representation, as here. Ihave expressed the proportionality rule by means of the D'Arcy Thompson_type system of transformed coordinates-an idiom horrowed from biologicalstudies of morphology. It is the existence of such mathematical!y generali:t.ablcproperties as these which lends visual coherence to the J'vlarquesan style.

There is a second feature of these small bone hair ornaments to which wecan draw attention. Because of their sm:111 dimensions, they are caryed, onthe whole, in vcry low relief. They therefore provide exccllent instances ofthe method through which J\1arquesan art transforms two-dimensional etuaimages into three-dimensional free-standing objects. The procedurc is simple,since il amounts to 'wrapping' a 'flat' filII/-form around a cylindrical anna­lurc. That is 10 say, whereas \\-'estern art 'adds a dimension' 1'0 conyert a two­dimensional flat figure (a drawing) into a three-dimensional figure (a work ofsculpture), the l\larquesan approach is rather to represent a two-dimensionalfigure in a three-dimensional 'space', while le:1ving its two-dimensionalityessentially intact (like the two-dimensionality of a postcr applied to a cylindricaltelephone-pole). The bone tiki is a 'curved' work in the nat, rather than a three-

riG. S. I ('!4, I~ji, The double lik! formin bone hair urnamCnL~ with hllkcnkrmzmalifun bad, 2nd ruw. SUlmr Von

den Stcinen, iii, ~K (abuve) and bL(helow)

riG. 8.1615. heillw. The h"kmkr,,,c'111otif: a "arianl uf ellil, via hupr <,chine.Source: Von den Steinen, ii. 1l)2,

illus. rllS and ISy

A: forehead to chinB: chin to hipsC: hips to feel

4

c. 8 em.

3

1.A<:B<:C2,A"B=C3.A>B=C4A>B>C

2

24 em.

A (forehead to chin) is approx. equal to A'width from ear to ear.

I

\/

~Tiki (post)160 cm.

B

c

FIG. 8.[6/6. The 11Ilcs or hod)· proportions in ]\\arqueS~ll ~rt_ S,-,"rce: VOll den Stcincn, ii,iflus, 7{; iii, jm6; iii, ~\'2a (<;J; iii, ~KI<)~

A'b.IS9. "",KENKREUZ·SOfEMA

:Vlost of these small cvlindricnl bone ornaments arc distinguished by the verysummary treatment gi\:en to the body and especially the legs, by comparisonto Lhe head. There seems to be a relatively consistent rule of proportionalitydictating the head/body ratio which is built into the Marquesan style. If wetake it that any .Marquesan plastic representation of the body .is essentially amodified cylinder, the proportional rule decrees that as the ratIO b~tween thediameter and lhe heighl of the 'cylinder' approaches unity, the ratIO .betweenthe head and the body increases in favour of lhe head. The underlymg con­straint seems to be the need 10 keep the proportions of the head/face withinbounds; the distance between the brow and the chin must remain approxir~l­

atelv equal to the ear-to-ear measurement, or in other words the (ice must fit,mo;e or less, into a square. This rule applies independentl~; of the abso­lute dimensions of the 'cvlinder'. As Fig. P1.16/6 shows, relatively .c1ong~ted

M;lrquesan representation"s (r) of the body show ratio.s between the (hmensJ(~nsof the head (A), the body (U), and the legs (C) whlch conform to the r~l[lOS

found in nalUre A< 11 < C. Reduction in overall height produced progressl\'e1y

432

l')G. 8.lill. "larql.lcs;,m artistic proliferation: fan handles di'iokd inlomultiple sqITleltts in difien:OI orK.'TIlarioru>. F..:aeh fan ~ndlc measuresapproximately zo ems. SQtifU: Yoollen Stcinm, iii, j)!\l and lJl\ of

S'1'It: alit! Gul/url:

Alternatively, bodies may face inwards but the fal..'t.'S outwards, as in Fig, 8.17/1(3)· i\nOlher kind of switching affects the orientation of figur~ in the \enicalaxi., (see Fig. 8.17/2), some of which Illay be upside-down relati\-c to the others.Finally, these fan hamlll.'S demonstrate a variety of 'distortions' of the headlface. If repTl.'Senred spanning the hroad side ofthe handle (as in Fig. 8.1711 (-.))the face flattens OUI to a shape similar to Illata Iwata. But if [he nccd is tor~present the tace projecting out of the top or bottom of the handle, making akmd of 'pommel', a completely new morphology is generated, which has notattooing analogue (Fig. 8.17/2). Steinen places these fan handles in:1 distinctclass, describing them as 'anchor' or 'doublc-.lnchor' handles. The nell' mor­phology of the head which is generated in these 'pommel' positions resemhles(from the side) Ihe beak of a bird or perhaps thc pointed nose and jaw ofa rep­tile. From thc frolll, howcver, the resemblance disappears, and one sees only aformulaic nose and mouth attached to a kind of llartcned proboscis which hasno ob\'ious counterpart in any family or creatures whatsoc\er.

S1j1tt: Illd Cut/urI:202

, Though this ch~l't", is prim~rill' cOI1<wlIcd with morphology and style, it might h,: as wcllto pointout, cSJlCci~Jlr in COnll"CI;'lll with lhese ornament' of can'ed bonc. th;\! ,\b,qHe'~11 WOt~S of ar~ arcalmost allla,'s crcated lrom nl~tcria1.1 "hich werc ,ignilicanr (i.('. r;tu,lIl)' pOlent))11 rhe'r ()\\Il I'lghl.These hail' ~rnamel\l'. fur ;n.t,mcc, we,,", m~Jc uf human bone; ,h,,;r pr01C"t;\C powers slcmm~d fr~1ll1

thdr SUbllau(( os much as their form. ,.\lrholOgh they arC inJependcm ohjects. rhey arC p.rts uf bod,es,both bccmse the' wcrc atl"c1,C(ltu Ihc body "f the wearer, .nd heclu..c riley "crc extraclcd f,om thebodies of olllc",,'... bone.< "'h;eh we.... sobsequenll) car\"e<l "nd made inlO orn.melllS_ Tlte> ..... th.usmuch more li~e t:ltt"" mOlifs th.n might be readih- apl"rem. Ta'too mmif. arc an>lorts mad,' of 1,,_onK flesh; 'h""" a....•an"'nrb made of bone>. hut Ih,' impliclti"n" in either cas<:, arc "c,,- much Ihe same.

8.17. Allll1;plirat;oll, TrtlIlSpos;t;OIl, aud ProIJlJscis-Formal;o,,:Fall Hamill'S

These small bone l.."'an ings are the most 'cylindric.!l' of anefacts. The whale­i,or'l fan h:mdles, where l'vlarquesan artistry is particular!} well displayed, arelonger and flatter, and ,Ire usually divideJ up into :1 series of segments, eachcarved with Janus figures in various dispositions (Fig. R.I7(1) e-.!ch of whichbroadly resembles one of the small eylindric.!1 figurines we ha\'c just heen con­sidering. Ilowever, each of these Janus figures is treated as a unit, so they arcinscribed in relati\·ely c1onbr:lted cylinders and hence have fully representedbodies and lower limhs. The fan handles present various poilUS of interest,which they share with the stilt-steps I will consider in a moment. They show,first of all, the tendency tow,lrds sheer prolifer>ltion which is such a markedfeature ofMaHluesan an. Although Lhey are relatively small objects (about 20

em.), each has a minimum of four figures represented on it, and often oneor even two more on the hase, besides sundry additional incised motifs, all ofwhich can be lraced back to the basic rlua motif. The proliferation seen here isa function of the rilual importance of fan handla-, which were important itemsof regalia owned by t.he mosl po\\crful chiefs and ehiefC'i.'>CS. A chief's powerwas a function of how Illany 'supportl..'TS' he had: chiefs engaged in deadlystruggles with onc another to increase their numerical follo\\ ings. These mul­titudinous fan handles refer symbolically to the salience of 'numbers' in .Mar­quesan politiC2l1ifc (Sl..'C Gel! 1993).

A second fe-.l.lure of these fan handles is the tendenc) for figures to switchtheir orientation; that is, he subjected to rotation hori:7.0ntally or vertically.This kind of switching can mean that both of a pair ofJanus figures may facethe same way and that consequently the back of onc of them is exposed as inthe second figure in Fig. S.17!l, bUI (his is alw:l)s offsel by the upper pair ofJanus fib'"Ures which will f:lce in opposite direclions, protecting th:lt side.

dimensional form. This approach to three-dimensionality via the imposition ofcun'ature onto basically fbt motifs is a diagnostic feature of J\hrquesan plasticart, "isible particularly in the rendition of thc faci:ll fe-JlUres ofanthmpomorphicfigures.>

204 Style (Iud C/lII/ll"e Style and Cli/iure 205

rIG.8.dl,jL A simple form ufstilt-stepwith rudim~ntary Elee at base. Source: Vonden Stcincn, iii, ~H

8,18, The Same Continued: Stilt-Steps

The process of proboscis-forrniltion, and various other geometric transforma_tions, can he perspicuously studied in lhe carVing on stilt-steps, a very charac_teristic form of~larquesan art (Steinen, ii, illus. 104-9; iii, PF, G, H, I). Thegame of fighting on stilts was a major sporting activity in lhe .I\tIarquesas, :-lndmuch c:lrc was devoted LO the carving and ornamentation of the hardwoodfoot-supports which were lashed to the main shafts of the stilts. All carvedfigures on stilt-feet essentially function as caryatids, supporting the projectingfoot rest. ·fhe caryatid figure rejoins the shaft of the stilt at an acute angle, cre­ating a triangular prism which can be subjected to ornamental treatment.Because outward-facing caryatid figures arc joined at the back to the shaft ofthe stilt, they are not carved in Janus form, like the figures on fan handles.Stilt-steps in thcir simplest form are supported by a single outward-facingcaryatid, as in Fig. 8.18/r. However, given that these figures were adjuncts to

FIG. 8.1712,jiJr kji. Doublc­an"hur fan handlcs withmorphulogically adaptedhcad. Soura: Von dcnStcincn, ii. '60, ilIlls. qS

rIG. S, 17h, left. Fly-whiskh,mdle from the AustnllIslands with similarproho,eis facc. Soune:nritish ~'fuseum i\l.\IS45S

Representing the face as a proboscis of this kind is a .feature of Po~ynesianart from olher islands besides the Marquesils; parallels wIth the fly-whisk han­dles imponed into Tahiti from the /\ustral Islands immediately spring to mind(Fig. 8.1713). This device occurs elsewhere in !\ttar4uesan art, i~ car b?rers andin stilt-steps (see below). How is it generaled? In effect, the. bndge o~ the n~lse

(which is hardly even suggested in the typically flat-faced Images of the kmdwe have been looking at up till now) is sharply seized and drawn forwards anddownwards, as if it were made of plasticine. The lips remain attached to theouter surface of the proboscis, at the tip, but al1 resemblance to :1 normalhuman face is lost'. None the less, I do not think the .Marquesan artlst~ actu­allv intended that these proboscis-faced creatures should be seen as, bIrds oranimals, despite their radical conrravention of the nonn~l ht~man 'fac~ sch~tT1a.They are simply geometrically reconfigured fac:s whIch fill spac: m the r~­quired way (e.g. accommodale the need to provlde fan handles wIth roundedpommels).

FIG. 8.1815, "boc'c. Stilt-t;tep "ilhrol~t~d duplicat~ figure. SOllra:\')ll den Steinen, ii. "53, illus. 249

FIG. 8.18/4, leji. Stilt-step withmultiple- figures and 'mooning'buttods. Source: Von denSteincn, J3G (7 and 8)

Style (Jwl Culture 207

of magically pOCent f.1ciaJ features (large OUTWard-facing eyes, especiallv) in aspace (an acute-angled prism) whose geometry is totally unlike that or a realhead. The reign,ing anti~natu~alism of ,~tarqllesan art has no trouble meetingsuch demands, mdeed, lIlgcnlOus solutIons to spatial problems seem fO havebeen pursued for their own sake. This spatial 'wit' is even more apparent in thecontonio.ns introduced into the basic cuyatid pose seen in our simplcst ex­ample: Fig. H. 18/4 SU~llll;lrizcs the variety of these poses ofsingle and multipleC'J.ry~tld figures .on sult-steps. The first transformation (of a single caryatid)conSISts ofrotatmg the body ISO degrees while lea\·ing the face in irs oricinalorientation, facing outwards. This device gi\'es prominence to the buttocks ofrhe re\ersed caryatid, which poke out, as ifit were 'mooning', to use the cur­rent expres..'iion. This butfock-poking gesture was I.:enainlv used b.,,- the "'laorito insult their enemies and ir seems prohable to me that ~larques~ns had thesame idea, in which case this device can be interpreted as an insult offered tothe stilt-fighter's sporting opponent. It is certainly more common on stilt-stepsthan anywhere else, though it is also seen on fan handles. The next possibletransformation is from one figure to two figures. This can be achie\'ed in twoways. First, by converting the single carratid into a Janus-form carYatid b\duplicating it and rotating by 90 degrees, so that {he nyo caryatids ~ow fac~forwards and backwards (Fig, 8.18/5), Or the I.:aryatids can be doubled by plac­ing one on lOp of the other (Fig. 8,18/6).

FIG. K.18/3.ltft.Transfomlatiun ufnose/mouth roproboscis

Fl(;. R.I8fz,jilr lift.From rudimentary faeelu l'Ofl"lpl<1c secondfigurl:: slih-slCp.Sflllrrt: \ un tlenSieincn, iii. ~G (5)

:~-'-'I '_. • • ..:. .;

-.:- .- - ~I : ,

-~

Style lind Culture

Fig. R,IS/3 shows the coordinate. transfonna:ions in~'olved in the 11':,Hlsfor­mati;)'n uf nose/mouth into probosCIS from a senes of stilts. It should he no.tedthat these arc nut diachronic or evolmionary transitions, Mar~lllcsan .u't!stswere free to prodlll.:e facial proportions as they pleased; pro~osclS f.\c~s nc'l1lyfill the small triangular prism left beneath the 'feet' of the lllal~ Car~,.ltld ~bov:;

'mechanical' raTher than evolutionary reasons explain this (hstortlon.' SIll:C Iris ;) means of making maximum usc of the available surface for the dclllle.ltlllns

a dangel'llus sport in which supernatural assistancc was most necessary, thetendency towards proliferation rapidly asserts itself, Even in the simple aT.ldrelatively early (180-1-) example J ha\'e just cited, a secondary face makes ItSappearance.u the base, below the feet of the main caryatid figure. V,ery com­monly, though, a second complclc figure is inserted beneath .lhe m;u~, upperone (Fig. 8. T8/2). Very many further complications of the baSIc I.:aryatld figurearc possible, which I will dt.-scribe in a moment. Before doing so, le~ me returnto rhe subject of the 'proboscis face' which commonly occurs on !,;tlh-steps.

206

FIG. II. I8{6._~la,il1l:dl)"

proliferatedstill-stell" i,height figures.Sourrt; \ on drnSteincn. ii. I]],

il1us_ 100) SCHEM"

209

FIG.8.19/1. U'u warrior'sdub with pro,cetl\·c facepr<Jlift:r:llion

Slyl" flllJ Culf/Jrt

8.l9. U'u: The Ultimate Double-Double-Double Tiki

Conflicts among the .\1arquesans were not confined to sports like stilt-fighting,they were also very prone to more serious barrles. Their main weapons of Warwere, ilS usually among Polynesians, heavy wooden clubs. The design of Mar­qucsan clubs was derived from the paddles they used lO propel their canoes,and indeed the non-chiefly combatants' weapon would be onc of these paddles,or a heavy version ofone for fighting usc. Richer and more important warriorshad clubs of a special and rather standardized kind, with carved decoration,called u'J/.

208 S~)'/( UIIJ Culfurt

Finally, the design can be clabonted by introducing further subsidiar~ fig­ure<> in various orientations. I made the point earlier, with respect to fan handles,that the theme of 'numbers of supporters' was of gn.-al significance in thesymbolism of .Marquesan chielly rebralia, hence the prolifcra~ion of su?sidiaryfigures. This applies ,\lso 10 stilt-steps, though even more literally, Since thefigures c,lrved on these arre~ac~s do indee.d. 'supp~rt' thei~ ~wnets clevate~position in the world. Hence It IS not surprising to hnd ~ubsld~ar~ supporterssprouting from the basic caryatid form. Onc way of domg thIS l~ by convc~t­

ing the 'limbs' of the main caryatid into lillie r'llIa fi~~res, that IS, the de.vl~eof hicrarchization mentioned earlier. The most stnking example of dus IS

anah'sed in Fig. 8.18/6. In rhis stilt-step, the main caryatid is doubled with asub;idiary 'supporting' one, below (with a proboscis face). The h<.'ad of theupper caryatid is secondarily doubled twice more (these would be raised 'ar~ls'

were this figure not equipped with ordinary arms as well). The lower ~,ort1on

of the btxlv of this caryatid is represented by two r'tlUI figures as (hOrizontal,inw:lrds-pc;inring) 'legs; and its 'mooning' bUllOCks hne become two m.ore eflla

'faces'. $0 in total, this stih-step contains eight distinct tUtti figures to mcrcaseiL<; owner's chances of viCTory; or sixteen, since this is only one of a pair.

SteineR de,"otcd an enormous amount of attention to these carved clubs.which arc among the most impressive and commonest IVlarquesan artefactsin collections. The Marquesan club is a J:miform rilla image rendered as alethal weapon. As might be anticipated in the light of the previous discussionofbad-to-l'ront symmetry and the Marquesan form of split representation, thedub has no back or front, hence no 'unguarded' rear, and always shows itsapotropaic 'facc' to the viewer. The club manifests many of the stylistic char­acteristics of Marqucsan art we have already described. In effect, Ihe problemposed by the dub [rom the Marquesan point of "iew. is to incorporate refer­ences to the dUD schema inro an artefact whose overall form remains modelledon the smooth, ovoid, and featurel/,."SS canoe-paddle from which it is derived.Steincn shows how this is done in an analytical figure which I reproduce inFig. 8.19/1 (Steinen, iii, ~ ~ r). Destcnding, we scc, finH of all, a small protect­ive rlull face un the crown, thcll the upper part of a vcry large 'face' with twostaring cyes, whose pupils are themselves subsidiary faces. Below this, there isa hea\')' horizontal bar. This is sct ofT from the 'face' above it by a notch abO\'eand another indentation below. This bar constitutes the 'shoulder' line of theclub as an anthropomorphic figure. In the middle of the bar there is a funherprotruding (ilia head/face. Descending further, we see a symmetrical eomb­like pattern on either side connected by a 'V' shape, and below this, a protect­ive devi<.:e of the hllJmllacllz type. I shall shonly givc Steinen's ingeniousexplanations for these. Then there come two morc 'eyes' and finally a decorat­ive band containing e/fla and 'woven' motif". Steincn's other drawing ofa clubshows the depression of 'saddle' bct\\"een the faces on either side; there can beno doubt that clubs arc classic Janiform or 'double tjJ:i' images as well as per­fectly practil..'al weapons. Howc\"cr, Steinen's interpretation of the parts of theclub goes funher than this, since he is able to argue, com.'ctly I believe, thateach club is a wmposite orfuur, rather than two, main figures.

Stcinen's ;lrgument starts out from a very different-looking artef;\Ct on avery diHcrent scale, a small bUlle elua fig'ure which has a splayed subsidiaryfigure protecting its back, above which is a further hakml!reuz etull pro-­tecting the head. As with the 'mooning' figures on the stilt-steps we have justlooked at, the body of the littlc 'pi~'-back' figure is orientated forwards,showing the buttocks, and onl)' the head is rotated backwards (Fig. 8.19/2).Steincn's ingenious suggestion is that this is the basic layout seen in the club.The head which protrudes in the centre of the 'shoulder-bar' is the backwards­pointing head of a piggy-back rider. The comb-like patterns to the left andright below the bar arc the SplOlycd hands of the rider, and the 'V' shapebetween its shoulders, beneath which is placed rhe protel:tivc Iwh:lIkrellzmotif. The legs of the piggy-back rider are not seen, but its mooning buttocksarc-Ihcy hne becn transmogrified into the SCl:ond sct of ey<."S abo\-e the dee­orati\-e oond, at the bottom.

8.20. Fusion: Tlte Nlirralive Art ofEar-Omal1tenlS

Finall~, to round off this analysis of the ,\Iarquesan style, I shall briefly discussa less ferocious art fonn, indeed the most playful of the traditional i\larqucsan

FlO. 8_19/2.fu,lcft. ThefilllJrative modelfur V'I< Jisplayingsul>sjdiar~" figureon Ihe bad:.Suunr; \'on denSlcinen, ii. 163.il1us. Ijl

2JJ

"·10.8.19/."lrft.V'u exemplifyingSldnen's modelwilh subsidiaryhe~d prolccli'lg'the bad: of theb3ck' of Ihe club.S"unc; Von denSirinen, ii. 163.iIIu~. 151

Style and Cultu.re

In other words, the subsidiary figure whose head we notice between the OUI­

stretched arms of the large figure facing us, is a piggy-back rider, protectingthe hack of the large figure on the other side of the elub. .\Jot only is the backof the club 'proteeted'---by being symmetrical with the fronl-the bark o[,j,(ba,k ~f th~ c1~b is prutl.'cted by a subsidiary figure on the [rOllt (Fig_ 8.19/3)­At this pomt II must seem that ~hrquesan spatial thinking is simply runningout o~ dimensions in which to anticipate, and ward ofT, spiritual dangers.EspeCially when we remember that the back of the piggy-back rider is itselfpruteeted by a 'third-order' protective figure in the form of the 'hakenkrcllz'etua. At the same time, the interpretation which would see the subsidiary headand rhe buttocks/eyes below as simple rcduplications of the etua motif ~n thcfront of the club is not by any means ruled out; one could sa\, that the main'face' of the club is duplicated in the vcrtical axis (top to botto~) as \\ ell as theproximal-distal axis (back to from). However, it seems to me that the basicschema is exactly as Steinen claims, and this geometric pla\' with dimensionsis ahsolutely in tunc with the basic stylistic tendency of Ma;qucsan plastic arl.

SfJ'lr and CU/tll"210

FIG. 8.20{l. !'-<ar-plug withswing IUrl'Jlin' mo,·jngIOW:lrdS the dua!mafn hOil/Q

mutif. Sourt~ Von denSreinen, Ii. 1.j.8. iIlus. 128

Finally, let me turn til the second set of 'narralivc' ear-ornamenls, whichrelate to an episode in the myth of Kae (discussed in Gell 1993: 18+-6). Kaemarrie.<; Hina, a woman who rejuvenatcs herself by periodically being smashedin the sea, like a crab, before being washed up on the beach with a new, softskin, which hardens so th,\t she can become a young girl again. However Hinaand her companions do not knoll' how to give birth. The midwives on theirisland are two felll,lle divinities of evil omen, Fanaua, who deliver babies bycutting the mother open, so Ihat she dics. Kae teaches Hina and her womenhow to deliver babies properly and lhe FanaU;1 depart in anger (complicationsof childbirth were atlributed 10 their malign efforts). f'anaua really means'fused together double' thus, ,I banana which grows (as tht:y sometimes do)as two bananas fused together on one stem is a 'fanaua' banana (Dordillon,'fanaua'). This obstetrical myth is a vcry suitable one for narration in the form

Style and ('II/flirt 213

The mythic episodes referred to in the two types of 'narrative' ear-ornamcntswhich havc survived both relate to feminine themes and I think that thesc Orna­ments were specifically items of female attire. The first of these types is the'Pahuatiti's daughters tin their swing' motif (Fig, 8.20/1). Pahuatiti's twodaughters wcre persuaded by thc chiefToaetini to humiliate the mythic heroAkaui, by pissing into his kilva. They did so by swinging above him on theirswing, and just at the critical moment, directing their urine into his cup, as hcwas raising it to his lips. For (his affront, thc branch from which their swingwas suspended was broken off, and they were cast imo the abyss. (For a newanalysis of swings and swing-myths in Polynesia, see Lamndcs, ]Ollnwl ofthi:Polynesi(UI Society.) The unforrunate conclusion of the Slorv is not referred toin the ear-ornamcnts, only the delicious episode of the swing, which is held upby two 'sla\'cs' rdther than the branch of a tree. From a stvlistic point of viewwhat is interesting aboul this set of ear-ornaments is (he virtuosity with whichMarquesan can'ers manage to cram four figun..'S (2 sla\'CS + 2 girls) infO such atiny compass, often by inverting the girls, who 'swing' upside down (nos. Iiand 18). However, the design often transfonns itself into a purely formal excr­cise in rows of miniature hcads/limbs, as in no. 25. Altemati"ely, the designsuccumbs to the gra\'itational pull of the etua motif (perhaps because of itsintrinsic symmelry). J\'umbcr 33 is well on the way to this destination, whileFig. 8.2012 retains only the fainlest reference to the 'narrative' design, and hasbecome an t:luulmala Ililtil design.•

11 18

~~'~~

Style and Culture

,

212

Fr(i. 8.20/1. Narrn(ivc arr: 'I'ahuatili's daughter. on their swing' J"pic(cd on cat-plugs. Source:Von den Steincn, iii, 13 R

lQ 20 21

~. ~•• ~Q'Ci. )." , '

'" - --'-, , - -

arts if one excludes string-figures. These are the carved ivory car-ornaments(inserted into hlrge holes in the earlobes) worn by both sexes and discussedby Stcincn under the appropriate heading of 'miniaturplastik' (Steincn, Ii.136-48). As one might expect, mallY of these I.>ar-ornamcnts feature sundryellUl-type figures, reduplicaTed in various configuralions. ]Jut much more inter­esting [han these arc IWO types of car-ornamcnts which dcpan radically fromthe rcst of rhe art we have been considering in that they involve explicit refer­ences {O details of mythic narr-.ui\'cs. The only Olher comparable instance is theTauoo motif discussed earlier, va; 0 kma, which refers to a particular episodein the ranooing myth of the hero Kenol (for discussion, see Gell 1993: 186-8).

The fact that these ear-ornaments embody 'narrative' is, I think, connectedto the fact thai thc)' arc specifically l'ar-ornaments, and il is through the earthai myths enler Ihe body. The ear is, needless to sa)', an orifice of the body,and, as such, "ulncrnble and in need of being protected, But the main way ofprotecting Ihe ear was by I:luooing a special C:ltegory of l'tUil designs on thccheek, near the margin of the t..'ar orifice, Since the ear orifice was protectedin this wa" the ear-ornaments could function, for once, as purely 'secular'.,adornment. In fact, they might be seen as the plastic-art equi";l.lents of SonyWalkman headphones, continually relaying mythic mformation into the ears towhich they were attached.

Slyle IIl1d CUIIUft' S~ylc {lilt! Gil/Iliff 215

8.21. Conclusion: Cohel'('fIcl1 ill MaJ"qul1.wlI Ar/and Social Relations

It is time to review the problem of .~tyle in the tight of the ground just tra4versed. Style, J argue, is 'relations betwcen relations' of forms. The aim of thepreceding analyses of J\'1arqucs.111 art forms was 10 show how individual motifs(particularly in tauooing, but also in plastic art forms) could he transformedinto one another by \'arious mlXlifications. The Nlarqucsan style, from a fonn.11point of ....iew, is the complex formed by thc relationships which hold betweenall these transfomlations or modifications. That is 10 say, the constraints go\'­erning the production (inno\"ation within culrur.tlly prescribed parameters ofstyle) of Marquesan art\\orks were constrainL" go\'erning the possibility oftransforming a motif or form into related forms; onh if such a transfomlationis possible can a motif or form be said til 'belong' t~ the \'larqucsan style. Inthis sensc, the ~'larquesan style is both unitary and dynamic; it is a field of pos­sible or legitimate moti\"ic transformations, rather than the rurality of existinginstantiations of Such transformations.

The question that now arises is the relationship between style in i\larquesanvisual an and J\1arqucsan 'cuhure' in the wider sense. There arc realh twoquestions here whieh netd to be distinguished; first, the relationship be~\\TCn'culture' in the form of artefacL" and 'culture' in other guises-kinship, eC()­nomy, politics. and religion. SL'Condl}, there is the problem of the relationshipbcn\'een culture and visual-art st}Je, which is a much narrower question. Therelationship between culture and the material artefact production is at leaSTrelatively independent of stylistic considcrations, in that artefacts in a ,-arie~­

of st}'les might subscn'e the culrur.ll role assigned to artefacL<; in any givenculture. Lct me gi\'e an example, In the f'.'1arqutsan context, there is a demon­srrable relaTionship betw(.'Cn the general form of the kin."hip sysrem (Dra\'idian,with preferential matrilateral crnss-(:ousin marriage) and tattooing/body arts.Tattooing is a form of'wrapping' (pnhll til'i, 'wrapping in images') and pallll isalso a kinship category 11lC'Jning 'malrikin + affi.ncs' (Gell 1993: 176 fT.). ]nother words. rhe characteristic involution of the Dravidian kinship universe,where political succour antI brides come from an enclosing circle of matrilat­eral rel:Hives. is connected, via a scheme transfer, to bod" arts which enclosethe illdi vidual in a protective wrapping of t:lttooing. Lel ~s suppose, however,that Marquesan tattooing did not have lhe visual appearance it actually docshavc, and that it looked instcad like Samoan tattooing. Samoan tattoo motifsare generated in a quite different way from Marqucsant:llloo motif.", and rep­rcsent a different srylistic syStem altogether. But if, as our gedankcxperimentsupposes, the Marques,l11s had h.ld 'Samoan' tattooing in placc ofrheir own, itwould still have been perfectly possible IU adv.mce e:\3ctly [he same argumentconcerning the connection between Dnlvidian kinship and tattooing.

For this reason, onc has 10 bell afe of making inferences about the relation­ship between 'style' and clillurc if these inferences arc really based on the rotc

d

h

- •

•

g

DeB A

b•

A 8 C D

~•

ABC D

,~.~~m~.$1~~

of car-ornaments, if we are correct in believing that women were lhe ones whowore this t)'pe. And it must be said that nowhere docs Marqucsan artistry seemmore inspired than in thel";c microscopic tableaux of entangled bodies. Thereare five protagonists, whom it is impossible to identify with the charaCters inthe myth, though Sleinen gives a detailed analysis (sec Fig. 8.2013). The bird­headed individual at the bottom right seems to be the child being born, thoughthis may not be right bc<:ause in fact Marqucsan mOlhers g;l\'e birth supportedbv their husbands and lying on a platform composed of the bodies of theiraffinal/maternal kin. So this could be one of them. The iconographic interpre­tation is anyway less important than the tL"(:hnical exploit:nion of the idC'.l ofambiguously fused bodies, which is what the Fanaua myth is really about. Thisplayful and I think essentially non-apotropaic image is Lhe antiLhesis of the sep­aration and insulation of body from body which Marqucsan ritual rules aboutlaPIl (contagious sacredness) upheld with such stringency; instC'.ld, it showsbodies merging into one anmher in an indistinguishable tangle. R:l.lher thanbeing a ritual image in itself, it is an ironie commentary on a world suffusedwith the magical danger ofcontagion and unboundedness. But this raises issueswhich nL'Cd to be discussed in a more general cultural framework, The time hascome (0 leave aside these minute analyses of specific Alarques:m artworks andtheir transformations, and to raise once more the problem of the relationshipbetween culture and m'le alluded 10 earlier. It is possible LO do Lhis because, bynow, the formal and m~rphologieal aspects of the ."'brquesan style, in bmh thcgraphic and plastic modes, hnc heen more thoroughly elucidated.

FIG. l:I.20IJ. Ear-plugs depi~ting the obsTeTri~ myth displayin['( a fusion of bo<lies. SUlmr: Vunden Sicincn. ii, 142, illus. 120

216 Szvk IIlId Culture Style ulld Culilire 217

of artefacts in culture, quite indepcndently of their style. Similarly, one hasnot demonstrated a link between art style and culture if the arguments usedrely on iconography. This applies, for instance, to images of religious or cultsignificance. An image which represents or embodies a certain divinity mighttake on a wide variety of difTerent visual forms; there is no saying, a priori, thatthe stylistically standardized form actually assumed by a representation of adivinity is necessarily connected to any other cultural aspect of this divinity.The standard representation of the divinity might be different, yet every othercultur:ll parameter connected with the divinity remain the same.

In f:lCt, ir is an error to imagine thar 'culture' in some general sense, isresponsible for the visual style of arteEtcts. Culture may dietilte the practicalandlor symbolic significance ofartefacts, and their iconographic interpretation;but the only factor which governs the visual appearance of anefilcts is theirrelationship to other artefacts in the same style. Visual style is an autonumousdomain in the sense that it is only definable in terms of the relationshipsbetween artefacts and other artefacts; it is a mistake to think of 'culture' as akind of 'head office' which decrees, on the one hand, what form political com­petition will assume, and on the other, what artefacts will look like. Artefactsarc shaped in the 'intcr-artefactual domain', obeying thc immancnr injunctionsgoverning formal stylistic relationships among artefil.cts, not in response toexternal injunctions from some imaginary 'head office'.

If it is true, as [ have just argued, that the relative autonomy of visual styleimplies that the factors governing the appearance of artef.'lcts belong primarilyto the artefactual domain, then dues that imply a complete disjunction betweenstyle and culture? Kot quite, I think, but in order to formulate a theory of thisrelationship which docs not fall into the same pitfalls as Hanson's (1983: dis­cussed above) it is necessary to proceed with care. \Ve certainly cannot argue,as he does, directly from stylistic properties of artworks (such as bifold sym­metry) tu properties of socio-cultural systems. On the other hand, it may bepossible LO construct arguments connecting the 'axes of coherence' withinstyles as systems, and other systcmatic properties of culture.

This is admittedly rather a mouthful, so I must try to explain myse1fmoreclearly. What are these 'axes of coherence'? In the preceding discussion ofJ\Iarquesan art, I showed that relationships among artefacts could be producedby applying transformations; among these were:

plane geumetric transformations of whole or part-motifs (translation, rOla­tion, refleCTion, etc.)cutling-and-pasting operations (such as those described in the analysis ofhope vehille)coordinate transformations (rcconfigurlng motifs in different coordinatescf. rhe analysis of Ken or the analysis of 'proboscis formation')

• hierarchization (one motif being subordinated to another, e.g. a subordin­ate (II/Ii protecting the 'back' ofa superior elull as in the 'club' design)motif-transformation (one motifbecomes another, e.g. I'I{/(I ~ mala hoala)

transformation by figure-ground reversal (as in the analysis of shelldiadems)transformation from tlVO to three dimensions by front-to-back reduplica­tion (Janus figures)transformation from two to three dimensions by 'curvature' (small bonetiki)rule-governed transformations of proportions, as in the head-to-body ratiosolid-geometrv transformations (e.g. switching the orientation of the bodvwith respect t~ the head as in 'mooning' stilt-figures) .formalizarion (as with the elutl ----+ hakfllkrellz transformation)fusion (as in the analysis offal/aua car-ornaments).

This is quite a long list. My claim is that it more or less covers the typesof motivic or shape transformations which define the relationships betweenMarquesan artworks. 'fhis is the nearest one can get to an explicit descriptionof ' the i\hrquesan style'. It is reasonable to assume that no other ;lrt style, any­where, would produce exactly the same list if analysed from the same point ofview. Some items on the list might be included, but not all, and other trans­formational modes might be detected, which are not LO be found in Marquesanart.

The next step is to formulate the 11arquesan style in terms of 'relationsbetween relations'-g:iven that the primary inter-artefactual 'relations' arc theones identified in the preceding listing.

Ilere, it is import;mt to bear in mind that the 'list' has been arrived at viainduction from Marquesan artefilcts; it is by no means a theory abollt theintrinsic possibilitics uf inter-artefactual relationships. The approach I sug­gest is totally permissive in this respect. All I assert is that 'style' is founded onconnections between artefacts. This would apply tu 'the complete works ofRembrandt' just as much as to 'all Marquesan artefacts'. Relationships betweenRembrandt's works certainly exist, but they are nothing like the ones whichexist between lvlarquesan tattoo-motifs or carvings. Inter-Rembrandt relation­ships derive from handling, lighting, pose, physiognomy, etc.-considerationswhich are completely foreign to the _lVlarquesan collective style. 'Relationsbetween relations' (between Rembrandt paintings) would involve different fac­tors from the ones 1 am about to ad\'ance with respect to Marquesan artefacts,and would, accordingly, orient the discussion in an entirely different direction(emotional psychology, drama, etc.).

As it is, there is no difficulty in sensing a certain homogeneity in this list­ing, a certain basic trend. This homogeneity deri\'es, I think, from a structural

principle which one could call 'the principle uf feast diffirence', th:ll is to say lheforms taken by motifs and figures arc the ones ilH.;olving the lellSllllo(hjimlioll o[m:ig!JbQ/lrillg l//fIti[s (fJ"s;slml wilh 'he t'Jltlblisnmml o/n dis/il/clioll bell/leen (hell/.Thus, there is no intrinsic rt,.'aSOIl why a 'f."1CC' (milIa hoa/II) should look like a'st".lled ellffl'; but the Marqucsan style decrees that they should louk as mut:hlike one another as possible, given that they are, in facr, to be distinguished.Similarly, there is no intrinsic rca~m \\h~ a three-dimensional 'f:\cc' (on acan'cd image) should be a 'curved' version of a two-dimensional face l':J.thcrthan a fal.:c L':ln'cd in thrcc fully re-J.lized dimensions. BUI the Marquesan styledemands thai C"Jrved faL't.'S be represented as \:un-ed' two-dimensional faccs;that is, the least possible concession to three-dimensionalif)'. Simil;uly, tw~dimensional imabrcs become three---dimensional Janus figures b~ hack-to-frontreduplicJlion, minimizing the conceptual distance between two and threedimensions. A seated tl/Ul, to gi"e a further example, can be transformed intoa tortoise, by a purely 'virtual' shift in the canonic orienration of the motif(from \'crticallo horimntal) without any change in the form of the mOlif what­soc"er. The same considerations apply to coordinate transformations, figure­ground re\ ersals, and so on.

The principle of least difference also applies in the context of !/luarrhizatio1l.Marquesan art shares with other eastern Polynesian art styles a propensity forproducing what might be tconed 'fractal' figures, images such as the famedRurutan ';\'a' in which a figure is composcd oflcsscr figures (discussed abo\·c,j.II). Marquesan examples are provided by stilt-steps (Fig. 8.18/~) and the'dub' design (Fig. 8.19/1). 'Fractal' artworks exemplif)' the k':lst diITerenecprinciple in that (subsidiary) 'parts' of (superordinate) figures arc forms whicharc thcmsch'cs figures, and hence arc as similar as possible (in fact, identical)to neighbouring figures (or mOlifs) leading independent existences. Where,as is commonly the case (cf. Figs. 8.18/~, 8.1911, and 8.20/3), the subordinalemotif is aClually a transformed \'ersion of the superordinate one, p,lrts arerelated to wholes by the 'lc:'lst difference' principle as well.

The 'least difference principle' (cxcept possibly in the last instaIH.:e) is nOIdetectable in any specific Marquesan anefact, taken in isolation, bUT only in theellsemble of relations constituted by all Marquesan artefacts. 'Rehltions betweenrel:Hions' (bet ween ]\1arquesan artefacts) arc characterizell hy convergencetowards The 'Icast diffcrence' principle: this is the 'axis of coherence' of theiV1arquesilll style, overall. Since wc arc now considering:l IOTally :lhstmct prin­ciple which is noT concretely instantiatcd in any given Marqucsan artefact, butonly in the ensemble of their relations with one another, it may. be reaso~abl.e

LO suggest th:lt at this point we have arrived at a level of ahstractlon at WlllCh Ittn,ly be possible to pro\'ide a cultural intcrpretation of the .Marques~n style.That is to say, although it would be mistaken, as I argued earher, to behc\'e tI~aT

'l'vtarqucsan culture' dictates the visual form taken by Marq~es~n artef.lets (~lr­

et:dv il is not so unreasonable 10 infer that the abstrat:t pnnclplcs goveTlllllg- ,

'" Sly'" af/d Gil/iliff SO'{.: IIIII{ Culture 219

'relations between relations' belong to 'culture' in the wider sense, rather thanthe inter-artefat:lual domain.

Lct us consider the 'least diffcrem:c principle' from a cultural standpoint.Marquesan societ)', as I have described elsewhcre (II)I)J drawing partieularlvon the work of Thomas 11)1)0) was characterized by a~ acute preoccupatio;lwith 'social difference' in a political cuntext of 'devolved' or fractured hier­archy in which 'difference' was exceptionall) diffieuh to sustain, pT:lctically.Because of lhe pre\·ailing political instability and runnoil, and the vehemenceof starus-<:ompetition, Marquesans had conrinually to seek to reinforce andprotect their social identilies ag-.linsl external threats to their wealth power, ,and social support. This struggle to maintain integrity against the threat ofdis­persal and de-differentiation .....as expressed ritoally in an acute preoccupationwith rules of inter-individual conlact and commensality. Stringent rules wereobsclTed in social interaction so that an individual's lafrll (contagious sacred­nes...) should not harm others, or, by being infringed, harm its possessor. I haveshown in detail elsewhere how Marquesan tattoo and other body arts are con­sistent with this picture of continually threatened personal and spiritualintL-griry. But identity was only under threat huaus~ it was so labile, so relati\'eand situational. Marquesans L'Ould switch social identities at will (b\' name­exchange, adoption, and other praCtiCL'S) so we ha\'e the (apparently) paradox­ical situation that a culture whose cenl.ml preoccupation seems 10 haye beenthe assertion and protection I)f social/spiritual idemiT)" is also the culture inwhich identily is exceptionally tenuous. This wa... a society. 1'O recall our earlierdiscussion of Lcyi-Strauss, in \\ hich social Status was, theoretically speaking,wholly a maner of pedigree-and in \\hich it was, simultaneously, acceptedthat pedigrees could be arbitrary political fiCTions.

It seems al leaSI plausible to argue that there is a connection between the'principle of least difference' in i\larquCS:ln stylistic, and the prevailing preoc­cupation with differentiation in Ihe comext ofdissolution which is the hallmarkof Marqucsan socio-eultur,tl attitudes generally.

How did this eonson'lIlce arise? Here line must recall that the Marquesanstyle is only the sedimcnted product of an infinite number of tiny sot:ial initia­tiws taken by Marquesan artiSTS over a long period of historical dcvelopment.Each new artefilct, however slancbrdi",ed, cannot come into being without thenecd for stylistic decisions, be they ever so apparently trivial and inconse­quential. These stylistic decisions, from which the cohercnce, stability, <lndlong-term transformation of the Marquesan style ensued, were taken withoutdeliberate reflection, but never wilhout cogniz,l!1cc of a prevailing social con­text of social forms, pervndell by n dread of spiritual/political transgression.That is to say, there was .1lI e1eelive affinity between a modus operandi in theartefactual domain, which generated mOTifs from other motifs by interpolatingminuscule varintions. and a II/odlls op~rlll/{{i in the social rL':llm which Heated 'dif­ferences' arbitrarily ;lg';linsl a background of fusional sameness. The limitless

220 Style I/Ild Culture

fcnilit}' of the Marquesan style in generating variant forms, each subtly dis­tincl, (.'Oupled with its striking formal homogeneity simultaneously suggest anonn\'helming need to establish difference and a recognition of the merelyrtlalil'( character of all differences.

In this sense, it is true {Q say that the relationships among motifs and figuresin the Marquesan arl slyle arc akin rn rhe relations which exisLCd, on the socialplane, between the Marqucsans Themselves. Anworks arc like social agents, inthat they arc the outcome of social initiatives which reflect a specific, sociallyinculcated sensibility. This judgement coincides in essential rcspeers with theview taken by the Marquesans themselvcs of their :m. The motifs and figuresI havc discussed are all categoriz(.'<.I, in their language, as lil:i: 'images'. Thereare two other contexts in which this word has meaning. 'Tiki' (as the name ofa mythologiGlI hero) is the Marqucsan Adam, Humankind were originally cre­ated through the impregnation ofa woman made ofsand by the original ances­tor (Tiki), who produced a human daughter, who was in turn incestuouslyimpregnated by her father, who disguised himself by blal.:kening his face inorder to conceal his appearance. from this union humankind are supposed tospring (Stein en t(88). This maker and un-maker of appearances is Tiki, thefirst man. Here we observe the fundamental scheme-transfer between imagc­making and the making of persons.

The other meaning of ,,-k,- is 'portions' or 'shares' of some distributed object(Dordillon). If I cut a cake into t\.\elve sli(;(:8, each slice is a tiki of the cake.Human persons are tiki because their identities are defined in terms of the eol­ICl.:tivities they partiL;pate in and di\'ide from. At the same time, this usageadumbrates an idea which has been thematic throughout this chapter, namelythaI artworks are holographil.: fragments of the 'larger unities' to which they arcunited by stylisti<.: linkages. The riki (images/portions) which represent (ormore precisely, constitute) ctl/a, at'e holographic fragments, or refr:lclions, ofthe imaginary totality ofall etun. Artworks arc shares or portions of a tlistriburedIJbjea corresponding to all of the artworks in the Marqucsan system, dis­tribuled in time and space. This idt'J will be explored in more derail in rhe nextchaplcr.

9

Conclusion: The Extended Mind

9. I. Distributed Objuts

The discussion of Marquesan gTaphicjplasric stylistics, JUSt wneluded, hasbeen founded on the notion of a 'corpus' of artworks as a kind of spatio­temporally dispersed 'population'. Marqucsan art, considered as a whole, canbe conceptualized, macroscopically, as a 'disrributcd ObjL'CT' in time and space.Like the 216 separate ifems in a 24-place luxury set of china tea and dinner­ware, all Marquesan artworks belong to a kind of 'set', though, of course, amore loosely integrated set than the sets ofchina which arc presellted to youngcouples as wedding presents by rich relatives. A china dinner-set is boundtogether, as a distributed object (an object having m,lfly spatially separatedparts with diffcrent micro-hislOries) by prior design, that is, by the intentionalactions of the design and manufal.:turing staffofSpade, or Wedgwood, or who...ncr. The corpus of Marques:an arl, on the other hand, emanates from no suchcenlIal executive otg',mization, and has come into being only by historicalaccretion (and deletion) via a network of social relations, among Marqucsans(artists and patrons) and outsiders (collectors, scholars, elc.) over the courseof more than two centuries. Except, perhaps, from a stylistic point of view,Marqucsan art has only a tenuous unity as a distributed objec!. It consists ofno more than the detritus or exuviae of the onl.:e flourishing art-productionand I.:irculation system of the 1\tlarquesas now sundered and scattered, likethe bones of the Marquesans whose living hodies bore the tattoos which soimpressed \'isitors 10 the islands in the nineteenth century. All that remain arcmuseum specimens, curiosities in privatc hands, sketches and drawings, andscholarly teXTs, such as Steinen's (and this one). None lhe less, despite this geo­graphical scattering and contextual transformation Marqucsan art retains aninner integrity of its own, as a macroscopic whole rather than as an aggregateof fragments. Each piece, each motif, eal.:h line or groo\'C, speaks to every otherone. It is as if they bore kinship to one another, and could be positioned withina common genealogy, jus1 as their makers could be. Above all, eal.:h fragment ofl\1arquesan art resonates with every other, because each has passed, uniquely,through a Marquesan mind, and each was directed towards a 1\1arqucsan mind .

.l\llarquesan minds arc :md always were, of I.:ourse, minds belonging to indi­,'idual agents, different and distinct. I do nOl wanl to suggest that I\larquesanart is the product of a 'group mind' or collecti'-e consciousness. But in theensuing se<:fions I do want to approach, with due caution, the problem of the

222 COlldusiOI/: The Ex/ended ,Hllld Clillelllsilm: nl<' [x/elldl'd Alintl 223

relationship hetween the macroscopic characteristics of distributed objects(such as 'the corpus of Marquesan art') and 'the mind' in both the individualand collective sense. The pith of my argument is that there is iSlI!lJQrplty o!sfruc­IM( between the cognitive processes we know (from inside) as 'consciousness'and the sp:llio-tcmpora[ structures of distributed objects in the ancfactualrcalm--such as rhe Q'I/life of one parricular artist (Duchamp provides myexample) or the historical corpus of types of artworks (e.g. Maori meetinghouses, St:c below). In other words, rhe structures ofart history demonstrate anexternalized and collectivized cogniti\'e pron."Ss.

I must prepare my argument with care, for I recognize that I am traversingdangerous ground. Let me return to some of thc idem;; I introduced in earlierchapters, so th.at thc.<te may provide something by way ofa stable platform fromwhich 10 launch thc ideas I want (Q de\'e1op now. Thl."SC ideas, to reiterate,concern the structural isomorphy hctwl.'Cn something 'internal' (mind or eon­sciousncs.,,) and something 'external'-aggregates of artworks as 'distributedobjects' combining multiplicity and spati()--tcmporal dispersion with immanentcoherence.

The contr.l.st between 'internal' and 'external' will be familiar from preced­ing sections of this work-in particular sections 7.f)-I I, dealing with the cxtcr­nalist and intl.'fflalist strategies for animating idols. One major upshot of thatdiscussion, I hope, was that the rontrast bctwl."Cn 'inner' and 'outer' is alwa}sonly a relati\e rather than absolute difference. The contrast bctwl."Cn 'mind'(the internal pcrson) and the enernal person, though real, is only rclative. Ifwe seek to del\'e inside the person all wc seem (Q find are other persons-thehomunculi of Oennen-and if, as sociologists rather than as cognith-c psy­chologists, we try to give an ae<:ount of the external aspect of persons, wc findthat :111)' one social imlh'idual is the sum of their relations (distributed overbiographical time and space) with other persons (M. Strathern 1988; Gell'Str:nhernognuns' 1998). Our inner personhood SL'Cms to consist of replica­tions of what we are externally, as suggested in the parable of Peer Gynt andhis famous onion. So, hearing this in mind, it may nO! be so aberrant to sug­gest that whaT persons arc externally (and collectively) is a kind of enlargedreplication of what thcy arc imernally. Especially if, as I shall be doing, we con­sider 'persons' not as boundcd biological organisms, but use this label to applyto all the objects and/or evcnts in the milieu from which agency or personhoodcan be abducled.

Seen ill this light, a pcrson and a person's mind arc not confined 10 parTicu­lar spat io-temporal coordinates, but consist of a spread of biographical eventsand mcmories of cvenTs, and a dispersed category of material objeers, traces,and leavings, which can be :lItributed to a person and which, in aggregate, tes­tify to ;lgency .uHI paticmhood dming a biographical career which may, indeed,prolong ilsel f long aftcr biological death. The person is thus understoo(l as Ihesum tot.d of the indexes whieh testify, in life and subsequently, to the bio-

graphical existence of this or lhat individual. Personal agency, as inten'entionin the causal milieu, generates one of these 'distributed objects', that is, all thematerial diflerent:es ill 'the way things are' from which some particular agencyClIl be abducted. -

I recognize that this conception of personhood is both yague and abstruse.Fortunately it is oot Illy task to describe personhood as such, but only toabstract certain themes which can be brought to bear on much more d~rhdemarcated 'distributed objects' rhan [he notional 'object' which correspond~

to the aggregJte biographical effect wrought by the existence (rather [hannon-existence) of a particular agent or person. The relatively well-defineddistributcd objects testifying 10 agency that I ha,·c in mind, ~re, of course,categories of3rt objects.

The iUl."a of personhood being sprC"J.d around in time and space is a com­ponent of innumerable cuhu ....t1 institutions and practices. i\ncestral shrinestombs, memorials, ossu:l.ries, sacred sites, ell:. all h.a\·e to do with the extensio~of personhood beyond the confines of biological life ,'ia indexes distributed inthe milieu. The first example to which I shall deyote detailed considerationbelongs to this t.':1tegory, namely, the type of memorial carvings produel.-din northern I\ew Ireland, and some adjoining islands, known as J\1aIangan(Fig. 9. III), whose chata.l.1.eristics and significance haw been anal)'scd by SuzanneKuchler in a series of articles (Kuchler 1985. 1988, 1992). I ha\-e a particularreason for highlighting these memorial carvings because the\- instantiate par­ticularly clearly, not just the idea of 'distribution' (the obj;ct andlor p~rsonbeing distributed in time and space) but also the exrraordinar\', "ct essential,notion that images of something (a protOlYPC) are parIs of that -thing (as a dis­tributed object). This is Yrjo J-lirn's idea (SL'C above, Sect. 7.-+), traceable. as wesaw, to Epicurus and Lucretius-the idea that sensible, perceptible objects,gl\"C offpurlS of themselves-rinds or skins or \'apours-which diffuse out into[he ambiencc and arc incorporated by {he onlooker in the process of percep­tion. The purpose of Malangan, as we will SL"C, is rhe transmission of ancestralsocial effil."acy (social prestige, ritual privilegl.'S. land-rights, etc.) through thedisplay of memorial sculptures which are incorporaTed into successors as II/em­

ories (internalized visual images).

9.2. Malanga/!

There are some 5,000 c<lrvings of lhe type known as 1\blallg'J.n in collections,which makes them among the commoneST 'collcclible' ethnographic an objectsfrom <lny art-producing region of rhe globe. Yet Kiichler tells us that, eventhough they are still produced in \'arious fonns, :md arc to the highest dCI,'Teesalient in contemporary political and ritual life in New Ireland, hardly a singleone is actually to be seen ill sitll lhefl..'-thcy ;Ire all in the hands of foreigners,and, as physical objects, of no concern to their erstwhile makers. Being sold on'

224 CIJlldusion: The fXJended A1ind C(}lIdu.<i()lI: The Extended MultI 225

FIG. ').211. 1\labngan \.:arving. Source: nril;,;h_~lUSCUlll.1\CIl' Irdand registration no. ,884,7-28.1

for cash is the final 'death' of objects of this type, and has been since the latterdecades of the nineteenth century. Conceptually, from the New Ireland pointof view, a Malangan carving which has fulfilled its ritual role has rotted awayand is no more, and its furure as a museum piece is irrelevant. i\1alangan only'exist' as socially salient objects, for a very short period, during the mortuaryceremonies for important persons, during which they arc gradually imbuedwith life by being carved and painted, brought to ptrfection and displayed fora few hours at the culminating point of the mortuary ritual-unly to be 'killed'with gifts of shell-money. Once they have bGen 'killed' they no longer exisl as

ritual objects (which is why they may subsequently be sold to col1ecrors). Thegift of money which 'kills' the l\tlalangan entitles the donor to remember theimage on display, and it is this internalized memory of the image, parcelled Outamong the contributors to the ceremony, which constitutes the ceremonialassel----t:ntitling the possessor TO social privileges-which is transacted at themortuary ceremony and transmitted from the senior to the junior generations.

The i\1alangan, as an object whose physical existence can thus be measuredonly in days, or even hours, is an indcx of agency of an explicitly temporarynature, During the brief duration of the ceremony, the carving objectifies;dense ilOd enduring network of past and future relationships between membersof the land-occupying matrilineal units which constitute northern New Irelandsociety. Social relationships between land-occupying units are legitimized onthe basis of members' previous!V purchased righls Iii remember Afalanglin tanJillgs{II/{I //Iollfl, and thus to act as agents in perpetuating these motifs (in differentcombinations) in subsequent tv1alangan ceremonies, where these memories willagain be brieOy objectified in carvings (in varied combinations) and again trans­acted and parcelled out alllong participants, against ceremonial payments.

But let us consider the lvlalangan carving more doseh'. There' are variouskinds, but I shall discuss only rlle familiar painted wo~den variety seen inFig. 9. III, which take the form of ancestral figures accompanied by· a varietyof subsidiary motifs. Which panicular forms and motifs occur on any givenMalangan carving depends on the identities of the land-occupying unitsmounting the ceremony and thc particltlar strategies of political alliance theseunits anticipate for the future, once these alliances have heen ccremoniallylegitimized by pilrcelling out .Malangan memories.

The purpose ofa Malangan is to provide a 'body', or more precisely, a 'skin'for a recently deceased person of importance. On death, the agency of such aperson is in a dispersed state. In our terms, indexes of their agency abound, butare 1101 conccntrated anywhere in p:lrticular. The gardens and plantations ofthe deceased, scattered here and there, are still in production, their wealth isheld by \'arious exchange-partners, their houses are still standing, their wivesor husbands are still married to them, and so on. The process of making thecarving coincides with the process of reorganization and adjustment throughwhich local society adjusts to the subtraction of the deceased from active par­ticipation in political and productive life. The gardens are harvested, the housesdecay and become, in turn, particularly productive fields, and so on. That is tosay, aU the dispersed 'social effecti\'encss' of the deceased, the difference theymade to how things were, gradually becomes an objectifi:lble quantity, some­thing to which a single material index may be atrached and from which thisaccumulated effecti\'eness may be abduct~d. This is wl;allhe Malangan is; ~kind of body which accumulates, like a charged baltery, the potential energy ofthe deceased dispersed in the life-world. KUchler (1992) speaks of the carvingas :1 temporary repository f01" the 'life-force' of the deceased, but we should

126 COIlr!mioll: TIlt EXJelliieJ Alind CQl/dliSioll: Th( t:.wfl/dl'd tl1iml 227

pcrh:lps observe thai there is 110 ditference between 'Iifc' itself and this lifc­force; the [ife-force which accumulates in the J"lalanbran carving is the ncrresult or product of a lifetime's activity in the social world, not a species ofmysric-JI energy distinguishable l--:ltcgorically from ordinary life and activit}.

The mech,mism for the accumulation of dispersed life into the physil:alindex of the .l\Jlalangan c-J.rving is via rhe mechanism of fire. The process ofmaking the image is conceptualized as the building up of a fire from ashes, toglowing embers, to the final blaze (1t)9z: IO.J). The raw wood is charged withefficacy by a Icdmiquc of heating and burning (this is connected to the tech­nicaluse ofHre 10 creatc the vcry numerous and complicated holes and cavitiesin the finished carving. especially in the epoch before metal tools were avail­able). As the forms emerge (the (;anTf, a hired specialist, is guided hy dreamsscnt h\ the ancestors), the carving grows (conceptually) hotter and hortcr. Theculmi~ating pfl)(.;CSS is painting, at which time the carving is ready 10 redis­tribute irs accumulated charge or 'hear' at the climactic ceremony during whichit will be publicly displayed and remembered b~ those privileged to do so, andduring which it will, irself, 'die' and become cold and rotten.

The can'ing, as I mentioned, is understood to be a 'skin' for the dece;L'ied.The concept of a skin is of the utmost importance here. for a number ofre"dsons. In northern I'\ew Ireland, 'skin' stands for affinal relations. Politicalrelationships (primarily, control O\-cr land) arc founded on strategic affinalrelationships created by ties of 'skin'-skin stands for the transQrrohlt ptrson,the person di"ided up, recombined, and reconstitllled. (For more on 'skin' cf.Gell '993: 23fT.) The can,ing, as Kuchler implies, is both a thn:e....dimensional.solid wooden 'container' for ancestral life-force, but at the same time. as anexternal surface (a two-dimensional field) it is a parchment on which particip­ants in the Malang<1O ceremony inscribe their anticipated affinal alliances, in theform ofspl:cific painted decorations in red, white, and black. Jt is the transactedmemory of the external, painted form of the can'ing which \,'jIJ legitimizefuture relalionships between land-occupying units. As a carving/container,the Malang-m is OJ repository of past 'social eITecliveness' accumul:ned andcontained, while as a spectacle, an exterior, the Malangan projects The futurethat these P:lst relationships will produce, as a result of the legitimizationof cert:lin anticipated relationships (between affines) that the Mahtngan cere­mony enables. The Mahmgan, in other words, mediatcs and tf:ll1smits agencybetween past .md future. Though the carving itself exists only within certain(restricted) time-space coordinates, conccptually, it is a fCII/POrtl/(y dispersedobjet'f, .Ul object III no specific time or place, bur moving through time and

place, like a thunderstorm. . . .But the notion of 'skin' has an additional slgmficance, which may not ha\'e

escaped the reader. Let us proceed to the climactic moment, at which the par­ticipants in the ceremony witness lhe can'iog in its final form an(1 register itsmemory. At this moment, the pri\ileged ones (lhe ones who h;i\e 1l1.\de lhe

appropriate ritual p:\Ylllents) receive 'the knowledge of Malangan '-that is, theright l'O reproduce not just the visible form of Malang-.m but the social rela­tions, including land-rights, which this visible form indexes. 'fhe word whichis used to refer to this empowering knowledge is WII/If.', which, among otherthings, means 'smoke'. Since the carving has, at this stage, a surface charged tobursting with fire and heat, it is in no way puzzling to find that the idiom usedto describe the transmission of potency from the can'ing (index) to the recipi_ent (spectator) is a fiery One. But we also find, irresistibly recalled to mimi the- -,words of Lucretius cited carlier (SL'Ct. 7.4) in which he associates the 'flyingsimulacra' of things. the 'idols' which are emined from things and which enableus (Q perceive them. with '[Iinle bodies] in a state of loose diffusion, like smokewhich logs ofoak, hcal and fires emit'. "Ine Epicurean model implies that imagesof things are diffusible parts of things, jlL<it as smoke is a difflL',ible part ofburn­ing logs. Lucretius' association between the emission of heat and smoke­formless, quasi-material diffusion from the objccr-and the more familiar 'idols'which haye \'isible form is reC'J.pitulated, almost magically, in I\ew Irelandthought. For rvU1It (as powerful knowledge) i'l not just the 'smoke' which eman­ates from the fiery surface of the can'ing; equally it is 'likeness' (the simulacra)in ,-isible form, the 'skin' of the cacving which is internalized a.<; a memory image.

The 1\-lalangan carving is a skin....idol. which like the 'gossamer coats ofcicadas' is distributed in quasi-material form in the memories of the onlookers,who internalize the ancestral 'skin' as a new 'skin' of their own, a new skinwhich anticipates new 'skin' relationships with affinal parmcrs. J\lemorizingthe image is a way of growing a new skin inJemollJ' and thus projecting a newidentity into the future. Poor Peer Gym (above, Sect. 7.11) could only acquirehis new skins by undergoing all the manifold vicissitudes related in Ibsen'splay, resulting in the biogruphical accumulation of memory-skins which he dis­assembles along with his onion. New Icelanders proceed differently, for theyhave elaborated the art of transacting 'memories' as a conscious, public, strat­egy; the accumulation of memories is institutionalized rather than being theproduct of happenstance, like Peer Gym's. Their accumulated, interlockingmcmories consist of intcrn.l1 skins, medialed via the Malangan, which can betaken apart and reconfigured at will.

This happens :11 the climactic moment of transmission, as the surface of theMalangan, animated by fire, is dissip:l1cd in rhe form of Lucretian simulacrawhich are internalized, more or less as internal body parts, by the privilegcdonlookers, who, in this way, rcceive the substance-not just of the ancestralbody, but the entire agel/live ('(Ipt/city of the deceased-for future redeploy­mcm. This is, as il were, the supreme <1bduction of agency from the index,in ['hal the other's agency is nOI jllst slIffircd via the index; it is also therebyperpetuaTed and reproduced. Thus memory becomes a socially engineeredmedium for the transmission ofthc power [0 change rhc world and shapc thecourse ofevcnts, rather than a mere passive reg-istration of the past. Once the

228 CundIlS;OIl; Tn( Ji:rlt:ntltd Milld 229

index has b<:cn .....irnessed in its 'charged' state, it is a mere corpse, dr.l.ined ofits power, l:x:cause whatcvcr memory images others may form of it (by wit­nessing it in its inert condition in a mu!'>eum or a shop) they will not be thosemcmories, the one!'> uniquely stipul:ltcd as potent and efficacious.

The example of Mahll1gan ;lft is useful in thal it can start to undermine thedistinction we commonly make between the marerial and the mental (or cog­nitive) with respectllJ material culture. The Malangan arc indisputably mater­ial objects, but the socially rdrmllt Malangan arc internalized images whichNew Irclanders carry about inside their heads. Being a material object is merelya transitional phase in the biography ofa Malangan, most of whose existence isas a memory trace, or, more idiomatically, as:m internal 'skin', The iVlalang:msof northern :"Jew Ireland irself-rather than the Malangans in collections-arewalking about, making gardens and political speeches, engaging in exchmgctransactions, marrying and having children, yet, paradoxi(,;all}', they arc notaccessible to external ethnographic observation at all. Only an extended surveyof past, present, and prospcctiw Malangan ceremonies, and the associated kin­ship and land tran...actions-which Kuchler is completing as I write-willreveal the ideal form of Malangan as a regional system of socially disuibutcdmemory Images.

We, meanwhile, cannot pursue this fascinating prospect, but we can con­tinue the general theme by turning our attention to similar systems of region­ally distributed artworks {this time, (,;onsisting or enduring objects, rather thanmemory images) forming a dynamic whole, by reterring LO certain well-knownstudies by Nancy Munn (1977, 1986).

Nancy Munn's work is particularly salient in the context or this discussion, inthat she has devoted p:lrticular attention in a series of studies, to the rela­tionship between material indexes-Gawan canoes and Kula valuahles-and(social) space-time. The argument on which I ~Im embarking turns on thisissue. My thesis i... that 'cognition', or more precisely, (,;onsciousness, is amental procL'SS through which subjective tempornlity is constituted via a pro­cess of transfOnn{/lioll of conscious experience over time. In lhe next section Ishall brieOy present Husserl's model or the mind as a series of ,modifica lions'of perceptual/memory images. Concurrently, I am arguing that the 'indexL'"Sof agency' which exist and circulate in the external social world create, so tospeak, an inter-indexical space-time field which bears an analogous structure,that is to say that it, [00, consists of a series of transrormations of contents(images) over time. This thesis will be instantiated later, with regard 10 .theworks of Marcel Duchamp and .~1aori meeting houseso But before we arc 10 aposition lO embark on this argument, it \\ould be helpful 10 consider funher

the qucstion of the relation bet \H.'Cn art\lorks (or other indexes of agency) andspace-Time, In this area Alunn has made certain '"ery important contributions.

Nancy JVlunn's anicle 'The Spatiotempornl Trnnsf(mnation of Gawa Canoes'(1977) traces the biography ofGawa canoes, which start life as trees growingon the land held by a partieu!;lr clan and are fashioned into canoes by membersof other elans (moving through exchange pathways internal to Gawa in theproces!'». Once made into C'.tnoes, they enter further exchange pathways withinGawa. Here they arc transacu:d against the yams which are trallsmiued fromwifc-gh ing matrilincs 10 wife-recei\Oing ones. So in a sense, the canoes ate con­verted into yarns on Gawa, while outside Gawa they arc com'erted into shellvaluables-which are exchanged for them by their e\·enrual users, men fromother islands in the so-<:alled 'Kula Ring' who usc them for carrying on (J\'er­seas (incer-island) exchanges.

From the point of viell' ofrhe Gawa matriline whid, has exchanged a seago­ing canoe for certain shell valuables (valuables that will themselves be cir­culated in overseas exchanges) lhe canoe is stiU 'owned'; but it is owncd inanother fonn, as shell \Oaluables rather than as a wooden canoe. Alunn detectsa consistent process of de-materialization herc. The canoe that was, heretofore,a heavy, rooted, earth} thing, a massive tn:e, becomes, b} dcgrees, somethingtorally immaterial, or rathet, mat.erial but unconfined, That is, the canoe is con­verted into a 'field of influence'. This field is generated by the magnetismexerted by the Kula-exchange valuables into which the canoe has been con­verted. By virtue of heing thc unencumbered property of the owning clan (atype of property designated as kilomn) these \·aluables have the power to mo\cother '·aluables (of different origin and type) in the direction of Ga\la, andrecipnx--ally, as they tta\-cl outwards, carrying the name and fame of Gawa farand wide. Munn argues that they are converted ultimately int'O what she calls'sociotemporal spacc-time'. This space-time is not SO much a dimensionalmanifold as a field of forces (like an electromagnetic field) exerted by objectsof value (indexes of agency) uhim:ltely attached to powerful persons but cir­culating in the milieu. This field constitutes transactional space polarized bythe multiple for(,;es generated by objl.'Cts in continuous motion and undergoingsuccessive metamorphoses. Each of these kitoum (unencumbered valuables) istraceable 10 a member of the o\\ning matriline, who constitufes its social pointof attachment, where it (X,"3.ses 10 be a liberated object and becomes a partiblecomponent of a person, its original hole, so to speak.

A Kula operator, one who participates in the inler-island .lIld imernalexch~lllges in KuLl valuables (arm-shells and necklaces) is ~l spatio-temporallyextended person. The actual mechanics of the Kula system havc bcen dis­cussed so often since MalinOl1 ski's original description (1922) that it is hardl)necessary to prO' ide a detailed account here. Suffice it to sa) that men \\ ho par­ticipate in the system do so lxx.-ause the) can lay claim !O ownership of1..;/011111,

that is, Kula yaluablcs II hich arc their own unencumbered ritual property, 1101

230 Glllulllsion: The f:.'xle/ldl!d Millll COlldm;MI: '!he Extelllj,.'d Mimi 231

valuables which they may be holding as intermediaries between one NlOum­holder and another. The relationship between kitoum-owncr and the valuablewhich is held is indissociable, as jfthe kitoU//1 were a body-part, but at the sametime the kilOum is an object thai can be transmiued ahroad as an exchangeitem, that can circulate freely in space and time, and that can be com'crledinto another obj<.'Ct. The anachment to the original o\\ner persist.<;, howc\'cr,because important kiwuni arc individually named, rCl..'Ognized, ObjL'CL'i, andwherever rhe kifumn lr.:Jvcls from island LO island, lhe name of its owner \\ illtravel with il as well. The l·;trnl/II-holdcr insens his kiNnon into one of theexchange 'palhw,tys' (kedll) to which he has access, in exchange for returngifts which do nnt match rhc hiOl/l1I itself, hut which are:1 sort of rent, p:lid bythe recipient, for the privilege of being the one LO serve as intermediary in theforwarding of this prestigious J..,I/Qum to its ultimate destination. Thus a flowof wcalth is generated in thc opposite dircction to the passage of rhe ki/(}/fIJl.E\'enrually, though, aftcr passing through many hands, in different Kula com­munities, the Itituum will cm:ounter anotht.'1' valuable in the sysT.cm \\ hichmatches it, being equal in renown. This will be another ltifoum, the unencum­bered property ofan equally important man on somc dist:mr island. When thishappens, the originalltifoum (an arm-shell, leT. us say) will 'marry' its opposite­number ItirOllm (which would be a Ilt."cklace, because an arm-shell can only beexchanged for a m:cklace and vice versa) and the necklace, ,~ill begin to tI'J\'elback towards the originalltilowlt-holder, again passing through many interme­diaries along: the way and setting up further countervailing flows of wealth.The aim of Kula operators is to gain control of numerous kiwunI, manipulat­ing the pathways of exchange so as to contrive that their 'names'-nttached toprestigious valuables-travel far and wide. A man whose name is known in di!>­tant communities as onc who controls the pathways along which renowned\'aluables are IT:lnsferred, can influence the calculations of Kula operators infaraway places. He is, so to speak, more than a merely incarnate man. lie is anexpanded and clisscminated being, present here, there, and everywhere beeausehis namc is attached to circulating objectx, and still morc because the mo\"C­mcnts of these objects nrc influenced. at long range, by his intentional agency,his calculati\'eness, and (magically assisted) persuasion (Munn 1986).

The Kula \'aluables which arc associated with a Kula operator's name areconceptualized as indexes of his boJily presence as .1 person of commandingpowers; from lhem, distant recipients abcluct not jusl his power but his bodilybeauty, for these Kula valuables are, after all, body-decorations as well as body­parts, and they arc regarded as be:lUtiful as well as :lllcient and prestigious. Asa distributed person. the Kula oper,ttor attracts weahh as a young man :lltraC~s

lovers; other's minds are swayed by his long-range allure and rokens of \hlSlove in the form of \'aluables, sp<.:ed towards him. ;\<Iunn (1983) tells liS thatthe ~nking sySlem of Kula \'alu:tblcs corresponcls to the ranking srslcn~ ofKula operators. The ranking scheme applied to valuables opposes neW-Ish,

non-prestigious, itcms with little of the goldcn patina that comes from years ofpolishing and handling, and which have yet TO be associah:tI with the names ofmany famous men, to ancie"" tr<.:asured itcms, which powerfully evoke theidentitit."S of men who, lhrough Kula, haw rranscended time and space, whoare timelessly pote",. attr:J.cli\·c, and influcntial. An important arm-shell ornecklace does not 'stand for' someone important, in a symbolic \\ay; to allintcnt.<; and purpuses it is an important person in th,lI age, influence, and some­thing like 'wisdom' inheres in its physil.."3l substance, in its smooth and patinatedsurfaces, just as they do in the mind and body of the man of renown to whomit was attach cd, and from whom it has flown away as :In idol of distributed per­sonhood. (Once again, we can draw an analogy bctwecn The Kula valuable aS:I

migratory bodily index and the flying simulacrll of Lucrerius.)Dut we cannot place the whole weight of the discussion on the 'distributed­

ness' of distribUl'ed personhood, :tt the expense of the core of agency whichlies at its heart, Ilow docs one, in practice, become a great Kula operator, aman able to 'mo\e minds' at great distances and dominate an expended regionof social spal..'e and time? How docs one become so enchantingly aHracti\e,so irrcsistibl) persuasive, that the paths of intcr-isl:llld exchange l..'tlO\crgeineluctably in thc desired direction? Only through knowledge, intelligencc, andcalculation. For success to accrue, the Kula opeI'JIOr must possess a superiorcapaeit} to eng-Jge in strategic action, which necessitates a comprehensiveilltemal model uf the extern,ll field within which Kula \'aluables move .Ibuut(et: Gell 1992(/: 280-5). The operator musr be able to comprehend the mani­fold and inordinaTely complex field of cxchanges, must be able to rememberinnumerable past histories of exchanges, and evaluate lheir outcomcs. He mustconstruct 'what if' scenarios lhat illlticipate the fUTure with precision, guidingstrategic illl.ervcntion. His mind, in other words, must work :IS a simulationdC\'ice-and this indeed is what all minds do, more or less-presenting a S) n­optic \'iew of the totality of Kula transactions, past, present, and to come.

In his 0\\ n person, the opef:,ltor must reconstruct a \\orking simulacrum-ad~ nantic space-time n1.1p of the maze of Kula transactions, so that, \\ith som­nambulistic dexterity, ht, knuws \\hich delicate strings to pull. E\erYlhingdepends on the l.."Qherence of inner strategic intentions grounded in accumulatede.'\pericncc and memory, and rhe historicall~ produced world 'out rherc'-thereal world in which minds, objectified in exchange objects, expand, meet, andcontend. The sllccessful Kula operator controls the world of Kula because hismind has become coeXTensive with that world. lie hns internalized its C:lllsaltexture as pan of his bcing as:l perSall and as an independent agent. 'Internal'(mental processes) and 'outside' (lrans<!t:tions in objectified personhood) hn\'efused together; mind and reality :1I"e one, and-not to PUl 100 finc a point onit---somcthing like godhead is achic\·able. This (relal ile) di\ inization throughthe fusing IOgcther of an e:\pandcd. objectified agenC), and the m~ riad C-,llIsaltexture ofthc real \\orld seems 10 me 10 he the ultimate object'i\e of Kula. h

suggests, III me at any rate, path\\a~s towards tr:tnscendcnce which arc asaccessible to us, secular souls and die-h.1rd matcrialists, as to the inhabitants ofthe Melanesian islands which participate in Kula trans..1etions, but that is onlyby the by, The point that J wish lO extract from all this is more limited. It issimply that when we come to consider the expanded, transactable. 'persons'and personhood on which the Kuhl. system is founded, we arc brought torecogni,.,c Ihal 'mimi' elm exist objectively a,~ well as subjecti\'e1y; that is, asa patrern of transactable objects-indexes of pcrwnhood, in this instance,arm-shells, and m:cklace5--1as well as a fleeting succession of 'thoughts', 'inten­lions', 'mental states', etc. The Kula system as a whole is aform ofcognition,which takes place outside the body. which is diffuscd in space and time, andwhich is carried on through the medium of physical indcxcs and transactionsin\'ol\'ing them.

9.4. The Artist's CEuvre as (J Distributed ObjecJ

Let us turn from the consideration ofOceanic instances of'distributcd object.'"and 'distributed personhood' to an example closer to home. We, in the West,are familiar \\ith one fonn of'distribllled objcct' (indexing a distributed per­son) above all--the Q'I/T;re or 'completc works' of famous ,mists. Any artist ofrenown is represented by numerous works, disseminated in various collections,and also capable of being reassembled for retrospective exhibitions, or pub­lished in a de luxe cdition with a complete catalogue raisO/llle.

Let us consider the characteristic make-up of an artist's fl:llrre, an artist ofthe kind with which \\ e are most familiar, a post-Renais.'\3ncc professional artistwith a distincti\"e personal style and a personal critical follo,,"ing. The artist'sfEuvre consisrs primarily of a series of finished works, produced, it ma~ be, atdifferent places, and subsequently distributcd in numerous collections. Thesefinished works arc usuallv dated or datable, and can be assign cd to a chronolog­ical sequence, early works, middle-period works, late works, and so on. So theal'list's O;'lIvre is bOlh spatially dispersed alld lemporall~' dispersed. ltfier theartist's death, once the tI!Ut're is complete, it constitutes, as it were, an in~epcnd­ent chunk of space-time, which L"3n be accessed \"ia each work indi~'ldually,

each standing, indexiL"3lly, for all of them and the historiL"3I-biographlC3l con­text of their production.

Uut thc artist's (J!/wre docs not consist exclusively of finished \\"Orks cach oneof which st,lnds as an independent entity. If we srudy the output of manyfillllOUS artists (e.g, Leonardo, Michelangelo, Constable) we fl.nd, that ~lU.~er~icall\' speaking the greater pan consists of 'preparatory sluthes for illS e

~ .[ h" ··cal vallieworks rather than finished works themselvcS. 1\ oreO\'cr, t e 11/st(//1, .,. , '" d' t roduced for theplaced on these ostenSlblr 'provISIonal tee mica stu ICS, no p

, . 'h d" t or even greater Ihan theart public bUI for prl\ate use III t c stu 10, IS as grc:.-a , _ .. '. . Ih '1' matter) From"alue plaeed on Ihe fimshed works (s:deroom prices arc ano e .

Condusillll: Th~ £xu"rll'I1 Mi",j III

an ~i~!oriC:l1 poim of view, th~ ~rep3ralllry sketches arc ill\'aluable, becausethey mform us about the cogmtl\'e prO(."t."S,';;c:.'S of generation of the finishedworks produced f?r public exhibition. Moreover, they arc Often crucial point­ers to the underly,?g tre~d of development of the artist's style, and indeed thed~velopmcnt of WIder historical trends in an (e.g. the relationship betweenConstable's sketches :mu later ninCleenth-cCflturv art including 11111> ' .', ) Th . . , rcssJOn_IS~, e~c. '. c a\'~ll~bilit)~ o.f sketches and provisional \<ersions of works allowsus ms~ghl l~IO artlstlc actmty as a process unfOlding o\'er (cognitivc and bio­graphical) time.

Meanwhile, the distinction I ha\'e just df"Jwn betwtX'n 'p,cparatory 'tud' ,d'li'hd I." ,Ies

an illS e wor"s IS not absolute. Because we know the dates of finishedworks, we arc also ablc 10 see these finisbed works as bcillfT simultaneousl', "'f' . ." \,prep,arat~ry. stU( les or later works. Thus, Cezanne's earlier 'bathers', a;d

cerram oj hiS la~ldscapcs, while conceived independently, actually serve asprep,~rato~y studies for us Gralldes Ba;glleuus-a wOl'k in which cezanne triesto epltOlmze, and fUrlher develop, a long series of prc\'ious experiments O\'ertwcnT)' or thirty years,

. l\:an~' artists produce works in recognizable series, consciously C\·oh'ing ad,st~ctl\'e trcatment of ~ p~rricu!armotifover the course of their c:."3reer. Braque,for. mstance, started p:untmg pICtures on the themc 'the artist's studio' (fea­tunng a canvas Oil an easel) from the 1930S onwards; this series culminates in~~e I950S a~d 1960s in a number ofunforgenabJc masterpieccs which synthe­sIze Braque ~ maturc style. \Ve also have famous 'series' from Picasso, Hacon,Monet, MatiSse, etc. In Other \\'ords, it is frequently thc case Ihat works of artfo~. 'm?ments' of ,lcmporal series, not JUSt lx"Cause they are datable objects(oflgmatlOg at certam space-time coordin.1tcs) but bcl"3Use Ihc" form lineatrer

they are ancestral to, and descended from, other works in th~ lE/lvre. Take~tog~~her, they form a maem....objeet, or lemporal object, which e\"Olves over time.

~rn~[]y, we may notice that the constitucnts of an artist's (J!/lvre do not justPOint ~orwards' ill time, as the 'preparatory Sketch' points upstream towards~he.finJshe~ ~vork, or, Baco.Il's first '~ope' I~ints towards his later 'Popeli'.

rtlsts also remember prc\'IOUS works In maJ.:.ing new ones, 'quote' themscJn:sand e\'en ,Produce downright copics and rcplicas of pre\'ious work. \Ve do no;kno'." whIch of the two \'crsiotlS of The Virgin ofth, Rorl!J by Leonardo (theonemLondonorth· . p ')' h .. ' ... e one In ans IS t e onglnal and \\ hleh IS the COP)' all thatcan be saId f, . h ho h " ,I or sure IS [at t (. lsplay the same degn..'C of technical exccJ-ence and both equall)' " 'L· .1.' A ' ", ' . '. , . .le conaruos. rusts are nowadays repnmanded by

CflllCS for repeating I'> ~n ' , . ,',' '" 'd '. • e 1SC \es slllce t liS IS eonSI ereu short-changlllg thepublic who dem d . . . - .'. an ConllllUous lIH10\'anon, Hut actually :llJ artlsttc practice islIle\"uabl)' dep' d h" ' ,. en ent on \\" 0 esa e repentlon, Ol herwise the concept of'stvle'(whIch depends on some degree of resemblance between all the works il; anlEuvre) would be impossible to appl~. Without repetition, art would lose itsmemo!'y. Indeed, the concept of a 'preparator) study' implies. in ilself, that

GfJndUl;QII: Th~ t:xlmdrJ Mind232

with the protcnl'ion!retention arrows looping over and under the row ofch~ono~ogu:~lIyarranged \\orks (dots), but thai \\ould have resulted in ratheran IlIeglble dl~gra~. The point or this diagram is simply [0 communicate the ideathat we ca~ l~lagllle the artist's a:uvre, at the macro-scale, as one indivisiblework, conSlstlllg or many physical indexes (wnrks) but amounting to ;\ single

, These .rel:tions, from n~w on, will be called strong and weak prolclltions forpr~eetlve o.r furure--onentared relations, and strong and weak relmtiOflS

for. retrospective' .or past-arientated relations. The reason for using theseparflc.ular terms wtll become clear once I introduce HusserJ's model of time­conSCIOusness, which I intend to do in a moment.

Refore.' ~o thai, though,. let. me present an ideal model, in summary form,of the arfl~.t .'0 (J:lIvre (as a dlStnbutcd obj<.'Ct) as a temporal-relational diagramor ~lap (hg. 9-4/1). What we have arc a spread of indi\'idual dated works~'hlch form the. nodal. points in a network of tcmporal relationships or ;Jroten­tlOn a~d retention. 511lce all these relationships arc temporal ones we couldha,'C dlsplayt.--d the indi\'idual works in Fig. 9.4/t simply as a linear ~equence;

CIJ1ldusion: The F.xlnlJeJ MmJ 235

F1G.9.4/i. The:lnist'$ fnlt·U:lS :I

dimibutcd object

•

weak proIentiorl (precursor)

strong prOlentiOn (sketch)

weak retention (/e:apitulalion)

End 01 career

• Slfong retentioll (copy)

1/dvd..IaI WOf1( at art {painbng, sketch,.<.,

Chronologlcal sequence

-----

KEY----~.--.

•

•

....._._- •.•._--~..-.-.... -- -- ..:;;,-- .............. It,,~._ ....

.~-. _"............. -.. - .

........ . ---"'" .• Ar.. ,.' •• -- • _.-:::.~_~ _-. •

11 • If ~ - -_.._. ~ • --... •V--. -ll / ......... ¥

• --. ................/. -",tr.----.'\. ~JI'~ -I ... "'" -...... ......... -----. • ...--. • .... tt. "': .

/" '---.............. - ~ .... .. -.-Start of C<lf8eI

234

the artist will suhsequently 'copy' his own pre\'ious work (in private, 'study'format) so a.~ to produce the subsequent 'public' work. So the mesh oftemlXlralconnections between the worb; in an artist's trllvre points in both direcrions.!\ny given work of art, in gross terms, considered in the context of irs maker's(l'l/vre, is likely to be boch a 'preparation' for later works, and a 'recapitulation'of prc\'ious works.

Howc' cr, where both artistic projection and artistic retrospection are atissue, we are able to render this rhought a little more pr<.'Cise. There is adistinction to be drawn berwcen, say, a 'prepararory study' which is producedin the process nf designing a subsequent work which is envisaged in concreteterms, and the rather weaker relationship between a 'pr<.'Cursory' work .....hich,undertaken as an end in itself, subsequently turns out to be ancestral to somelater work, which was not specijira/~v ellYisaged at the lime of its production.While, in practice, there may be a shading-(lffbetween the 'prepararory sketch'and rhe 'precursor' ofa subsequent work, they may, none the less, be opposedas ideal types. Similarly, we may oppose the 'artist's replica' (like the later \'er­sian of the Virgin o.(,j,e Rocks) which is produced with the idea of specificallyreplicating a previous work, with the unintentional replication of prcviousworks which is a n<.'Cessary feature of the origination of new works; that is,the painter, intending to produce new work, reproduces his previous work be­cause stylistic coherence and painterly praxis (deeply ingrained artistic habits)demand it. Once abrain, in reality, there is a shading-Qff between intentionalreplication and unintentional recapitulation, hut the ideal-typical distinctionmay be allowed to stand.

Thus, we may distinguish two relati\'ely 'strong' temporal relationshipshetw<."Cn works and two relatively 'weak' ones. The 'strong' relationship sub­sists, in the future, or prospective, orientation, between the 'preparatorysketch' and the finished work, while, in this orientation, the 'weak' relation ishctwecn the 'precursory' work (the first in a series which was not planntd ill(It/vance as a serics) and subscquent works in the same series. Conversely, in thepast, or relrospecti\'c orientation, the 'strong' relationship subsisL'i bet\vcen the(past) original and the subsequent COP)', which is intended ro replicale this pastwork, while the 'welk' rcl;ltionship subsists between the 'original' work which-through a process of stylistic evolution in which nOI everything changes allat once-is parlially rertlpillllaw/ in subsequent work .....hich returns to it,modifies it, and dnclops it further.

We now have the technical concepts in place to essay a general model of theartist's (euvrr as a distributed object-in particular, as a distributed object inlime. since the dist'riburion of the artist's (t!/lvre in space, though an intercst­ing historical question, is beside the point here. The elements of our modelarc the separate components of thc artist's total (£lIvre, thai is, the 'completeworks' as individual itcms, which arc mutually related via the four relationsjust described.

COlldllSioll: Thr EIrr"drd MlntJ

tcmpoml entity, like a persistent thunderstorm which is made up of many,quasi-instantaneous, flash(,"S of lightning. The artist's reI/wI' is an object which,so 10 sp(''al, is IlIl1de out of tim,; not the tenuous, dimensional time of physics(,\hieh I have discussed, and indeed defended, elsewhere; ef. Ge1l1992a) butthe kind ofsuhstantialized time which Betgson named til/reI'. Bergsonian duree,a model for which is provided by biological evolution regarded as a releonomicprocess rather lhan as a random ilccumulation of chance mutations, has nosignificance for the physicist, but that is not to say that tiuree has no psycho­10~"lcal or wgnitiye validity as a concept. IndCl..-d, there is every reason to thinkth;1t personhood, under.>tood cognitively, is coextensive with subjective temporalexperience. To refer to a person as a possessor of 'consciousness' is to refcr toa series of cognitions arranged tempor:Jlly along an axis of tiuret. But here were,lch the crux of the matter. The chrtlllologically arranged set of works whichcomprise an artist's «'lIvrl' arc a set of malerial objects; they are not a personor:l set of subjeethe cxperiences (cognitive states). They comprise a set ofindexL'S from which the artist's pcTSOnhood and :lgcncy can be abducted, as wasdescribed earlier (Sect. 9.2). Rut at the &1mC time we can easily conceive that'remembering' something which happened in the past is very tikI' 'copying' apicture that one h:lS painted in the p:lst, or that 'making a preliminary sketchfor a picture' is very /ik, mentally anticipating some future happening or eourse

of action.In other words, the ammgclllent of indi\'idual works in an artist's rCl/vre,

c.1ch of which is panly a recapimlation of previous works and panly an anti­cip:ltion of works as yet uncommenced, seems to generate the same kind ofrelationships hctween illt/cxes (which are objects in the extcrnal world) as cxistbetween 1//(lltol slates in Ihc cognitive process we recognize as consciousness.In other words, the temporal structure of index-lo-index relations in the art­ist's lEuvrr externalizes or objct:tifies the same rype of relations as exist bet weenthe artist's internal states ofmind as a being endowcd with consciousness. Theartist's «'lIvrr is artistic consciuusness (personhood in the cognitive, temporalsense) writ large and rendered public and accessible.

But where docs \:ognition' take placer-in the artiSt's head, or on his can­vases? Mostlv, of course, the cognitive processes of any mind, esp<.'cially o\'era whole biographical career, arc inaccessible private experiences which leaveonly the most indecipherable traces. I\nd we could hardly aggregate an artist's(£1I~rr :lS an unified 'temporal object' unless each of the individual worls had,al. one stage, originated :lS an ill1e1lliol1 in the anist's mind to produce suchand such an index (i.e. a state of mind giving rise to artistiC agenc.y). BUI t,hegencrate-and-test model of cn.-alive :lgency, which we b.rieOy examllled ~arl.,eron, reveals most de3fly that 'thinking' takes place outSide us as well ~ Insideus. The l>OCt writes down his lines, :lnd then scratches the~ Ollt, altenng ~ndimproving his \·erses in ways that crucially depel~d ?n t~e eXistence of physl(~_~traces of his previous (mental) uCliviry. And lhls IS sull truer of the graphiC

CondllSion: Th, Ex/mdd ",i,uI 237

anisl, who continually uses his own past production as a Spur to his futureproduction, altering and modifying freely as he gues along. And in more gen­eral terms, the artist li,'cs sumlUndL.J by his own works, completed or half­com~lete~, ~"hieh litter the studio and prm·ide him with an cver-present recordof hts activity over many years (maybe till his activity in the case of artistswhose works find no buyers).T~e a~tis{'s l£1wrr can be unde~rood like this; each separate work is a

~o<hficatlon, :l recension, of previous ones, The leftovers of a particular cycleIII a carecr-long gcncrate-and-test sequence. To he sure, this model is some­what ideali:zed, and much of art consists of routine output rather than lheresults of strivings after perfct:tion. But the more itl\'emive and historic­ally important arti..-rs do de\·e1up in this way, and their work can be read as acumulativc process of dis<:overy rather than the ll1ere exploitation of technicalprocedures learnt early, never forgotten, and never surpassed.

The concept that I want to draw out further from the 'generate-and-tcst'~odel of creativity is that of modification. For one ver.>ed in the philosophicalIItera.ture of co~nition this word rings bells, because it is a word llsed (inEnglish transl:mons) by Husserl, who assigns the concept of 'modification' (ofimages, perceptions, i.e. the objects of thought) a central role in his model ofconsciousness as a t'emporoll process. I have provided an account of Husserl'smodel of 'internal timc-<:Ollsciollsness' in a previous worl.:, and I propose to

recapitulate some of my comments on Husserl's model here. My reason fordoing so is that Husserl's model can help us to clarify certain f~rures of themodel of the artist's (EU"", as a temporal obj(,"Ct (or pe~haps one should say, asa 'trans-temporal object') that has heen presented in the preceding pages. Inpanicular, it will help us to escape from a serious contradiction that I have notyet brought to light.

Suppose we ha\·e two works, X and Y, such th:ll X is a 'weak protention'towards Y, that is that. X can be seen as a precursor ofY but not as a definite'study for' Y. Now, while X is in the process of being painted, Y is unimag­incd by the paintcr or by anybody else, in concrcte terms. An that one can sayis that the painter, while engaged on X, 'hopes' to paim future paintings, whichwill probably be related, technically and thematically, to X, but he only hasa vague intuition as 10 what his 'next' picture will be. Y is still (and only)a 'future' painting, nothing in thc world corresponds to it. Eventually, thepaimer will get round to painting his next picture (let us say, the next in aseries) and it' turns oU( to be Y. Now we have both X and Y whereas before weonly had X. On inspecting Y we (and the painter) are able to see that Y wasprefigured in X in very many respects, and, on this basis, we arc inclined to savthat Y recapitulates X, [hat Y is a 'weak retention' of X in that X is retained i~Y as a 'preliminary ,·ersion' ofY. Bm wait, there is a problem here, because wchave already supposed dml (in vague, if not specific, terms) Y has been protem/cdfrom X, that is, Y is visibly what X portcm!J' in terms orthe artist's development.

239

PRESENT(Perception),

-----', ---,~-: -= ~1--··---"-"" ,it I

Huss~rl, working on this problem towards the beginning of his career, whenhc was Illterested in describing cognitive proccsses in general so as to be able(eventually) to separate 'psychological' facts about (;ognition from transcend_ental philosophical certainties, put forward a most useful model of 'time­consciousness'. This is designed to depict thc systematic modification of the'noemara' (the objects of cognition) as a function of the passage of successive'now' moments, that is, shifts in the subject's time-perspecti"-e.

In order to expound his ideas, Husser! makes use of a diagram, of whichFig. 9-4/2 (from Gell 1992a) is a version. The horiZOntal line A~ B~ C~ D corresponds to the succession of evems or 'states of affairs' occurringat 'now' moments strung out between the past and the future. Suppose we areat 8: our perceptions up to date at 8. The tcmporal1andscapc at H consistsof the now-present perceptual experience of the state of affairs aT U plusretentions of A, as A', shading away into pastness. A' (A seen from B) is amodification of the original A-what A 'looks like' from 8 that is attenuatedor diminished, but still connected to the prescnt. Perha~s one ~n think ofthe 'modification' of A a.<; it sinks (diagonally to the left on Fig. 9.4/2) downinto the past (A ~ A'~ A"~ Am ... ) as a gradual fou ofwnsimilirudt>affecting the perceptions emertained at A as these are superseded by the per­ceptions entertained at B, C, 0, ctc. Our perceptions of the state of affairsas it is at anyone 'now' moment do not become inapplicable immediately, butonly grnduaJly, because the world does not change all at once and in allrespects. We can no longer, at B, say that the state of arTairs at A is 'now' thecase, because of the change of temporal perspective; but many of the featuresof A have coumerpans at n, olnd features of B holVC counterparts at C, andso on .

. Reten~ions can thus be conslrued as the background of out-of:'datc perccp­tlOns agamsr which more up-to-date perceptions arc projt.-cted, and significanttrends and changes are calibrated. As perceprions become more seriously Outof date, the)' diminish in saliencc and arc lost (0 view. We thus perceive lhepresent nOI as a knife-edge 'now' but as a temporally extended field within

PIG. 9.4/2. A versionof Ilusserl's diagfllm ofrimc-<:onsciousne!;s from

PAST --",..,.,_...,..., ... FUTURE Gel], Tlit A"thropolof,.Y (If(memory' Feed·forward from PBSt - future

{projectionl Tim;. SaUTft: A. Gcll 1992Q

238

I-low can it be simultaneously the case that Y is protenucd from X (as a nol­

'fcr-realized 'future' painring) and simultaneously thaI Y 'retains' X as a kind~f memory-trace now that Y has actually come into existence. How can weidentif" the 'Y' which was a proun/jon!rol1l X and the Y which is u rtltntion ofx, \Vh;" it would appear that tht:sc two Vs ha\'c cOOlr.ldiclory propertics; thatis can Y be a protcntion from X ami a retention of X, at one ami the samem'urnent? This seems to involve some kind of logil.":J.l conflict, yet, when wethink about the relations between works afan in an artist's U!Ui.,'re, we "",,,nt Lohave it both ways; we want to see later pictures 'prefigured' in earlier ones, andwe want LO sec 'traces' or memories of earlier pictures in later ones. These twotypes of rclation between temporally separated pictures arc. clearly noT thesame, yet when we place them together they seem to coll:1pse mLO one anotherin a most confusing way.

This if you like is the art-hiSlUrieal version of a familiar dilemma in thephiloso~hyof tim~the problem of evems and tenses. Let me explain this b.ymeans ofan example. Tomorrow I have a doctor's ap~"Ointment.1 portend thiSevent, today, as a future event which will (probably) transpire, but I do notknow, for instance, what the do(;tor will say or what treatment he will recom­mend. By the day after tomorrow, tomorrow's dtK.1:or's appointment will be apa.steve~t (of which I wiD have a memory, or retention) having transiently beena 'presen£' event (tomorrow). Obviously, this event (the appointment) is the'same event' whether today is 'today' (IS October 1996), or yesterday, or lastweek, or tomorrow, or the day after tomorrow, or whatever day you please. Yetthis unique cvenl, on thcse ~ari()us days, has the contradictory properties ofpresentness, pastness, or futurity; and how can 'one event' be past and presentand future without contradiction arising?, .

The answer to this puzzle is obviously that events like doctors' appolllt­ments whieh are anticipated in advance, become present, and fall away intothe pa~t do not possess attributes of pastness, presentness, and futurity in aonce-and-for-aJl way, but only transicmly, depending on the 'poim ohi~w'wehavc on the e"ent in question from a cenain 'now' moment, which continuallyshifts. The 'future' quality of a future evcnt ccrtainly colours our attitude to it(it has an irrtafis shading, to use the grammarians' term), bu~ ~vcn s~ w~ haveno difficuh\' in identifying a future event which was only 'annclpaled With theevent corr~sponding to our anticipations (more or less) which .ac.lUally doeshappen, and the memory of that event which becomes srreafts III a ~thcr

different way as it slips back into rhc past and becomes a '~ere. mC~lOry . Th.esame evcnt, as n possible future event, as a prescnT event wblch IS bemg: expen­enced, and ,1S a past event which can be recalled, remains one cvent, b.ut asour temporJI perspective on this event shifts,. t~le eve~l unde~goes a senes ~fmodijicaliulls from the standpoint of the cogn1tlve subJect. .It IS SLoen throu~hvarious thicknesses of future and past time, which alter Its appearance, Itstemporal palination, so lO speak..

240 CQnrll/siQn: The Extended MindCOlldmi/)ll: nil' I:.'xICIIII,,1! illillil

which trends emerge out of the patterns we discern in the successive updat­ings of perceptions relating to the proximate past, the next more proximatepast, and the next, and so on. This trend is projected into the future in theform of protentions, that is, anticipations of the pattern of updating ofcurrentperceptions which will be necessitated in the proximate future, the next mostproxim:ne future, and the next, in a manner symmetric with the pasl, but ininverse temporal order.

Let us continuc with Husserl's own explanation of his model. At B, A isrct:lined as A' (A seen through a certain thickness of time) and C is protcndcdas C, the favoured C<1ndidate as successor to B. Time passes, and C comesabout as C (presumably not quite as anticipated, but approximately so). B isnow retained in l.:onsciousness as B', related to (current) C as A' was to B whenB was current. BUI how is A related to C? From the standpoint of C, A is nolonger rcraincd as A', because this is to put A' and B' on a par with onc another,and fails to rcflect the fact thar when B (currently ll') was current, A was romthen only a retenlion (A'). Consequently, from the standpoint ofe, A has to beretained as a reUlltion of A', which is itself a retention of 1\: that is, as AU.

Ilusserl says that as A sinks to A' at H, AU at C, Am at D, and so on, a per­ception becomes a retcntion, then a retention nf a retention, then a retcntion ofa retcntion of a retention, and so on, until reaching the stage of final attenua­tion and sinking beneath the temporal horizon. The effect of this argumentis to abolish the hard-and-fast distim.-r:ion between the dynamic prescnt and thefixed and unchanging past. Past, present, and furore are all of a piece, and allequally dynamic in the Husser! model (embodying an important cognitivetruth) beC',lUSC any modifiC<1tion, anywhere in the system, Sel'i up (;orrclativemodifi(;ations c....crywhcre else in the system. Thus the modifiC<1tion in thc pre­sent which converts C into C automatically entrains corresponding modifiC<1­tions everywhere (B' --7 BU

, A"--7 Am, D'~ 0, cte.). 'The whole past sinksin a mass, taking all its arranged contents with it' (Findlay 1975: II). But thepast does not just 'sink' as the present pTIIgresses; it changes its significance,is e"'aluarcd in different ways, and sets up different patterns of protcntitlns,according t'O the way in which the present evolves. This dynamic past, andthc future which continually alters in complexion, cannot be accommodatedin 'physic;\l' time, but only in cognitive time. In providing his model of reten­tions, ponemions, mooifieations, etc., Husserl is not dcscribing an arcanephysical process which occurs to events as they loom out of the future, actu­ali%e themselves in the present, and sink into the past, but is describing theehanging spectrum of intentionalities linking the expcriencing subject and thepresent-focused 1V0rld which he experiences. 'Modification' is no! a ch:mgc inA itself, but a change in our view of A as the result of subset] uent acererions ofexpCT1cncc.

Husserl sumlllari~cs his view of internal time-consciousness in rhe followingpassage:

241~\'cry actual ~o\\' of consciolJsness is subject 10 the law of modification. It ehalll;cslOW the rcte~tlon of a rctenrion anu doc'S so continuously. There accordingl\' aris 'rcg~lar contrnuum of r.ctc~lion such thaI' e\'cry laler fIOillt is the rerenrion'of e\~:r~~rher .one. Ea~h rcten~lon IS alre,l(1)' a corllinuUl~I. A I(]~e begins and goes on sleadil)::ts nOll -pha.se dlangcs Into a was-phase, and our ImprCSSlonal consciousness constantl ...

llows mer Inro an e\er new rctentiona! consciousness Going dowo tho "' "" . . . . . ' ~sr~m,\~cncounrcr a COlllll\UOUS SCTlCS of rclenllons harkln" back to rhe starting po,"o' T hf"h . . . Z> • • oeacn sue retenflons a continuum or relcntlonal modifications is add~" d h" " """ If '. Ul, an t IS cnnrm_~um .IS.IL"C. a. rornl In th~ actuality Ihat is being relenlionall} prnjrt.'ted _.. Each relen.(Jon IS mtrms'call ... a COl\llllUOUS modification which so to say carn" 'h h" f". . . ' '. cs e CTlI'age n liSpa.SI In Itselr. II ,s not mcrd\' Ihe (:ase Iha! going uownSlTcam ~~h I" "". ., , , ......" car 'cr rClenunn ISconnnuously replaced by a I\C\\' one. Each later TClenrion is not mereh- , 000'"

od""· . ... - . muousm I ,canon Slemmmg from an on<r>nallmprcssion' it is al<'~' 000'" od"""'. D" • "" muous m I catIon

?f al! prevIOus OOlll'nuous modifications of the same sLarting roim. (1(}28: 390 citedIII Fmdlay 1975: (0) ,

Similarly,. future. C\"entfi, do not really 'change' as a result of the fact that,from our ~~nt of Vie.\\', t.hey ar~ becoming less indefinite, more imminent, andL~n be anl1Clpated With lO~rcaslllg degeCt.'S of prLusion as they approach. But,,,e ha"e a strong compulSIOn to view them in such a light. Hus.'>Crl's model~eats [his "ia a continuum of t."onlinua of prorentional modifiC<1tions. Proten­tlOns arc continuations of the present in the light ofthe kind of temporal wholerhe present seems to belong to; 'To be aware of a deYeloping whole incom­pletel}', and as it dc"e1ops, is )ct always to be aware of it as a whole: what is notre~ ~\Tiltcn in, is written in as yet to be wrilten in' (Findlay 1975: 9).

'. he As and Bs and Cs in Ilusserl's model corrcspond to 'e\'ents' or Slates ofaffaIrs. What I wam to suggest is t.h:u they can be replaced b\' indi,·idual works?fart a~ constituents ofan atlist 's 0!1I'1:f~. Thc.'>C are physical ~bjccts r.1ther thanevents, bur, all the ~:lme, they :Ire trares or indexes of e\'ellls, that is, the events

?r p~rform~nccswhIch brought them physically into being. What I am argu­m~ IS that If nne seeks 10 construct thc artist's (l'/lvrr as a unified temporalobject, the 5:tme basic 'law of modifications' applies. What this means is [harwe cannot see the artist's f1!tl'VrI! as a tcmporal object except on condition thatwe select one particular work as corresponding to a 'now' moment and sccall the other works in the f)1l/vr/! as eithcr 'past' or 'future' works in r~lation TOthe 'now' defined by rhe panicular work that we hl\\'c selected as our temporalvantage-point.

. Thus, to return [0 our prcvious discussion, wc can only see an earlier paint­mg, X, as a 'precursor' ofY (pl'Ote11lion ory from X) by situating ourselves ata vantage-point in til~1e IIf. mlticlt V do/!s /lot yc/ c.r/sf, and conversely, we canonly see Y as a recapJluJatl0n of X (a retelllion of X) by shiftin<>' our poinr ofvlmtage w a later 'now' :1l II'hich Y has come into existence, and X is a 'past'~\·(]rk. \Ve cannot occupy bOlh Oflhcsc \:ll1tage--poims at the same time. ThisIS the same as sa~'ing thai we cannot totalize an anist's rru<:re as :l temporal

1984) which popularized adl'anced mathem:llieal aod scientific thinking. Con­sequelllly. although the actual diffusioo of ideas might never be documeotable-and I certaioly do not intend to document it-rhe f.lct is that Duchamp wasprobably to some extent ;1\\ are, e"en if only indircctly, via the Cubists (aod thdrin-house philosopher, Princet) of the James-Bergson-Husserl conceplion oftemporal flux or lhe 'stream of consciousness'. So thcre might be an elementof rautology involved in using Duchamp ro illustrate a 'Husserlian' model ofart hislOry, when, in rolCt, Duchamp may ,lctually have set (Jul to illustrate il.

I Lowc\"er that may be, Duchamp certainly pro\;des by f.lr the most per­spicuous instance of an important artist "hose total U'urre repavs stud\" asa network of protentions and retentions fanning out from parti~ular w~rks

(particularly from his masterpiece La ,I,tlariie mise a lilt par les ciliboraires,mime---otherwisc: koown as !he Larg~ Glass (19Ir25)), I cannot, obviously,do more than hll1t at the richness of Duchamp s U'lIvre here-that \\ouldrequire a monograph to add to the many thar already exist on this artist,certain of which l:opiously document the basic idea I am exploring here (sec,especially, the work ofC F.. Adeock 1984).

Duchamp's \\ork is, cssentiall\", ahoul the notion of the LlJ1ltinuum in lhatit is based on r.he exploitat'ion of the idea of the 'fourth dimension'.This'dimen­sion, I should immediately sa}, is not 'time' in the ordinary sense, especiallynot time as a mere measure ofduration, or physicists' time. The 'fourth dimeo­sion' for Duchamp-as for cerrain of his contemporaries-was esscntialh the'real' but srrictly Imrepresmlab/~ domain beyond, or encompassing, the 'o~dio­ary' ",orld we Ii\'e in and percei\'e in the normal way. Duchamp's work ori­ginates in a mathematician's par.lble, A two-dimensional object (io the plane,e,g. in Abbot's famous nat/amI) casts a one-dimensional shadow. A three­dimensional objL'Cr (the kind of object we arc familiar with in our 3-d world)casts a two-dimensional shadow, So what kind of object would cast a three­dimensional shadow?--<lb\'iollsly, that would ha\e to be a 'fourth-dimensionalobject'; which is thus something which one ('''an conceive of, bur not represent,because to do so requires more dimensions than lI'e have at our disposal, here iothree-dimensional land . Duchamp's art, 10 simplify radically, consists of a seriesof esscntially cornie attempts \0 produce 'shadows' offourth-dimensional entit­ies, or at least to suggest procedures for oblaining these shadows of fourth­dimensional objects by e,~t!"lpolating the shadows of three-dimensional ones,

To begin with, Duehamp merely identified the fourth dimension vaguelywith a Symbolist never'-neve!' land. In 1910-1 I he produced 'Symbolist' workswhich culminate in the VOIII/If /vl/1/1 (I/ld Girt ill Spril/g (19J I)~ which we willeocounter in:l diHerent guise bier. Frotlllllid-l~)11 onwards, he fell under theinfluence of the Cubists, The Cubists were interested in the fourth dimension,not as a symbolic myth but as a 1:ICl of subjectivc experieoee, and under Cubistinfluence Du(;hamp's rlOlion or the fourlh dimension became aligned, moreand more, with Bergsoni.1n duree, The underlying intention behind the 'clas­sic' phase of Cubism was to cre.lle images which were 'realistic' in showing the

COllelllSioll: Th( Exullded A'fil/(I

objett which can be rq,rarded SIIh specie aetemitatis. All we carl do is compile a'tile' of dillcrent tempoml perspec(ives on the cellue as a whole.

Suppose an <trtist creates, in the course of his c<treer, 500 works which wecan number Op. l through to Op. 500. We can constrUl:t the artist's cellI:re'from the standpoint of Op. I' (i.e. as a very '-ague set of protentions fromOp, I), Ot from mid-cucer (protemions :md retemions from Op. 250), or from'last works' (retenrions from Op, 500 and protentions towards works whichwe can onl\' imagine, which the anist might h:l\'c completed had he livedlonger). Dc-pending on which opus number we take as our point of \";lllra.ge,I would argue, we obtain a different, unique, patterning of protentional andretentional relationships between works, and thus a different interpretation ofthe artist's <</It'r~. There is no absolute sense in which any gi\'en 'Iork canbe seen, either as a recapitulation of a pre\ iOlL" work, or as a pn.'Cursor of afuture one; the ensemble of an artist's works, strung our in time, constitutesa dvnamic uns(';lble entity' not a mere accumulation of datable artefacts. We

J , " ,

can only appreciate it by participating in irs unfolding life,The reader may object :ll this point; 'this is all \'ery well, but try as 1 may 1

find it hard to connect what you arc saying to the actual «UL'r~ of any artistwhose works llmow-Canaletto for instance'. And I would be obliged to admitthar Canaleno's cityscapes, admirable though they arc, hardly seem amenableto interpretation as psychological documents, as opposed to topographic.l1oocs. The model I have been ad\'ancing best applies to artisL" whose (l!lIrr~

embodies a high dcgn.-c ofconscious self-reference and coherent de\·elopment.I am far from claiming that the model just ad,'anced would be particularly lL<;cfulin all art-historical context<;, Howe"er, the model can L"asily be made to applyin at least some historical contexts, and it is 10 one of these I shall oow lOrn.

9,5, The (Euvrc o.FMarce! Ducha111p

In a sense, I am going to cheal. I-Iusserl's model of lime-<:onsciousness datesfrum a period in which problems of space-time, continuity, and lhe relationbetween physical reality and mental states were very much. 'in ,tl~e air'. Histreatment of the subject shows, markedly, the influence of WIlliam James,while another contemporary philosopher who tackled these pro?lems-:-andwho may have exercised some direct influence on the (;ourse 0/ art hIstory-W:lS Henri Bergson. The birth of analytical Cubism, the llppe.aral.lCe o~

Ilusserl's PS)I/.'lIology /.If Intcnltll Time COlIsril/lI.1'IlCSS, and the pU,bhc:lt'1on ofBergson's most widely read work CreliliDc E:volulilJl1 were almost s,llnultaneollsevents (between 1906 .1Ild H)07). The artist whose w~rk.1 am ~OHl.g to usc toillustnHe my thesis, Marcel Dllchamp, undcrwent hIS formalwe Illtellecl~l.d

experiences during precisely this period. Though Duchamp ne\"c.r studIedphilosophy or mathematics systelnatica~ly, he.rea~ily picked up th~ ~dea~ l~~atwere doing the rounds in intellectual CiTdes l~ hIS ?'outh. More~.vel" he n;.I(~and mastered a nUfllber of texrs (notahly by POlllcare and joufTret, sce Adcock

Ctil/rllISiOIl: 1'h,' t.'.\'uwfed Alilld 243

244 Cllllcl,,,ion: The Extended Mimi CO>trlusioll: The /;.'xllIIded Milld 245

spectator what the object l'etI!(), mils like, rather than mcrcly whal Ihe object'looked' like. The earlicr ninclcemh-century artists admired by the Cubists wcrethe 'rcalists', notably Courbet and CoroL The philistine public was bemusedby Cubist claims to Ihe elTecr that superior 'realism' was their goal, bur theCubists could (itc copious philosophical precedents for their project. Il is notphilosophically unprecedented to suggcst that there could be more 10 guitarsand bonles than their visible appearance, \'iewed instantancously :lnd from afixed point of \'iew. Translating !.his insight into artistic practicc was morcdifficuh. CCz.annc had shown the way, for cxample in some of his late depic­tions of .\10m Saintc-Victoirc, whicfl-----..as later photographic rL'SL':lrchcs at thespots at which cezanne set up his casel wcre [0 prove--showed more of MontSainre-Victoire than could ever actually be seen from any on~ of these \'amage­points. These landscapes depict, not any fi.xed appearance of .\ tont Sainrc­Victoirc, but CCzanne's interaction with this objeer 0\"Cf rime, as he mO\'edabout in its vicinity and absorbed each of it... \'aried aspects. Or, in other words,.\-lom Sainte-Victoire is re\'ealed as a prouss, a movemcnt of Jllri~, rather thanas a 'thing'. The weak point of Cubist 'multiple perspective' was that it couldeasily degenerate into a kind of painted cinema, in which successive 'frames' ofa moving objcct, or an objL"Ct S<.ocn b) a mo\·ing camern, were simply pasted oncon top of another. Cubist theoreticians, such a.~ Gleizes and Metzinger, sedu­lously sought to emphasize the idealistic element in Cubism, its search for thepict'Orial equivalent of'the absolute' rather than the facile e\·QC3tion of dVn::l.micmotion and change. The Italian Futurists, on the othcr hand, embraced 'cine­matism' heartily, sincc, unlike their Parisian Cubist colleagues, they were spe­cifically interested in movement and dynamic phenomena (e.g. 13occioni).

Uuehamp Ix.'<.-amc a Cubist rather late in the day, just as 'classical' Cubistaspirations were beginning to unravel, and, being of:l saturnine disposition, hejoined the movemcnI in order to indulge his increasing prcdilection for mockery,ralher than because he Was a true beliner. (There were personal reasons forDuchamp's misanthropy, which have been eXTensively disinterreu by his bio­graphers.) Luckily for him, one of his 'satirical' Cubist works, the famous NudeDescending a Stllirctm.' of 1912, establisheu his name as a leading avant-g:lrdeartist in the United States, where he subsequently secured lifelong patronage,though iT rcsulted in his expulsion from the 'official' Cubist movcmelll (for'Futurist' deviationism). The A'ude transparently employed 'cinematic' methodsand was, in fact, b<lscd directly on stop-motion photographs by E. Marcy andothers. Duchamp's intemionally comic picture was designed La demollstl';lle lhefact that while 'realism' remained the ultimate objective, the 'fourth dimensiml'could only, in the end, be physical time, and 'realistic' images (Cubist re;llistimages, lhal is 10 S<lY) would always reduce 10 chopped-up panial images of theobject pasted onto one another or strewn over the canvas, <is in the notoriolls NI/dr.

Duchamp was more ambitious; he sti11 wanted to represent the unl'eprcsentablcflux of'bcing' (to employ the Heideggel'ian term) bur without simply reducingthe mu ltiplicity and fullness of experience to a series of partial snapshots. lie

FIG. 9.5/" Th~ La,!~ Glass or1'11~ BriJ~ stnpp~d b{Jf~ by IUTBa~hd{m fi.Y1I by MarcelDuchamp_ Phibdelphia !\Iuscumof An

conseque~rlyembarked, from t9t3, on a long series of prcparalOry studies ofawork which could, truly, adumbrate rhe fOurth dimension. Because almostall ofDuehamp's work, from t913, is part of a single, coherent project, which~ubscque.ntly, after the (semi-)complelion of the LI/rge Class in 192:;, extendedlts~lf unt1~ the cl~se .of his careCr, his re/li;re is panicularly imeresting from ourP?m~ of vIew..It tS !tlerally the case that Duchamp's (£U'/)re consists of a singledtstnbuted object, III that each of Duchamp's separate works is a prepararionfor,?r a. development of, other works of his, and all may be traced, bv directo.r ClrCUllous pathways, to all the others. This was intentional and ~xplicit,

slllce Duehamp's b:1Si.c objective was to CI·e.liC a fourth-dimensional entity, andan cell~re such ns hts IS perhaps as close as we will evcr get to possessing suchan enTH)'.

Considel'atio~s ofspnce-and the patience ormy readers--oblige me to confineth~ demons~nlI10~ of r~e dwraCICl'istics of Duchamp's (t'lIvre as n temporalobject to a dtscusslOn of jUst one work, or rather, the wllvre 'seen from' just onework. The ~\'ork I have chos~n. is one of the numerous studies for the LargeG!ass(see 'Ptg. 9.S/I).ln facl, It IS a study for the 'Capillary Tubes' which drnw

246 CO/lrlusion: 'I'll/·l.'xlended /J,lilld Cundl/siOI/: Till.' EXlended MlIId 247

rTG.9.5Iz. The ,Ve/wurk o!Sluppagl.·, by I"brtcl Duchamjl. '© Succcs,ion Marcel Duchamjl,ADAGl', l'aris and DACS, Luntlun [<)98.

off 'Love Gasoline' from the 'Cemetery of Uniforms and Liveries' on the left­h:md side of the lower half of the Large Glass and feed it into the conical 'Sieves'in the centre. These tubes are no more than a minor detail of the Large Glassitself, yet Duchamp devotes an important study to them, indeed, more than one.

The work I shall discuss is known as The lv'e/UJIJrk o(S/oppages (no. 214 inSchwarz's complete catalogue; Fig. 9.5/2). IT is both an independent work anda preparatory study for the Large Gla5I. At first glance, it looks rather like amap of some railway-system, the main depot (at the lower right, or south-east)serving a number of branch-lines limning out to the west and branching ,Igainto the north. One sees little numbered symbols on the lines (which may be st;l­tions) and other symbols which might, possibly, be bridges or tunnels. Each'branch' seems to come to an end in a 'terminus' of some kind. If onc looks atthis 'transport system' sideways on, from the lower right, one can pcrccive itsrelation to the Capillary Tubes in the LarKe Glass, for here the system is repro­duced in 'pcrspective' projection (from this point of view), rather than 'm:lp'projection. In the Ltllge GlaH the Two-dimensional map shown in the Ne/1J!orkhas become a three-dimensional perspective view of The Capillary Tubes (hint­ing at the transition from a 3-d to a 4-d world hy providing an instancc of thetransitioll from a 2-d world to a 3-d one).

The Ne/mork uf S/OPPIlXfS is rhercf()J'c a protentioll tow,lrds (part of) theLarg~ G/as.l, even though the Ne/mork only appears there in transformed per­spectlve. AT the snme time the lv'etrvork is a retention of certain earlier works.In panic~1ar it re.capitulates directly a piece called The Three Stalldard Stop­pages which consists of three curved wooden templates which were used todraw the curved 'tracks' shown in the l\;'etwork.

One idea Duchamp was working on was that in the 'fourth dimension'events which seem to us like 'pure chance' would correspond to necessitV. An'arbitrary' length, or a 'random' curve would, in the fourth dimension, be ~ome­

thing really basic like 'one metre' or 'a perfectly straight line'. Accordingly,Ducha~p took ~hree one-metre lengths of string, allowed these to drop freelyonto sticky varmshed boards, and from these arbitrary curves he cut templatcs,called the Three Sliindard S/oppages, which would be the basic geometricalforms and units of measurement for 'fourth dimensional' use. The networkis obviously a 'strong retention' of these templates, just as it is a 'strong pro­tention' towards the Capillary Tubes.

But there is much more to it than this. If we inspect the IVetwol'k more care­fully, we observe that it is painted over something else. Duchamp has not useda fresh, pristine canvas, but has done his design-work for the Capillary Tubeson a reused callvus on which more than one image has alreadv been inscribed.In ('lct, the canvas has already been used twice, for apparently different pur­poses. Counting from the 'top' layer downwards, the l ...'etmork of S/lippagesconsists of:

1. The 'map' of the Netll!orJ:;2. A (quite faim) preliminary line-sketch for the Cll/ire layout of the LllIge

Glass, as Duchamp conceived it in 19IJ, before many details had beenworked out;

.1. A version of YOIIIIK /vlan alld Girl ill Spring~Duchamp's 'Symbolist'canvas of 19' J, his first major p:linting, dealing with the theme of 'initi­ation' and (possibly) incestuous longing (for his sister, Suzanne, to whomhe gave it as a wedding present); cf. Pig. 9.5h.

:-.Jow obviously, Ouchamp could perfectly well have afforded to lise a freshcanvas (or simply a large sheet of paper) for both, and 2, if nil he required wassomething drawable-upon to do his design-work; especially if he intended to

sell his sketches to patrons once they had served their purpose, as most artistsdo and as he himself subsequently did. Instead, he produced what amountsto a series of works while using only 0111.' canvas; self-defeating parsimony onewould have thought~which he made up fur later by producing, or havingothers produce for him, many identical replicas of his earlier works.

Clearly, therc must be more to ['his ["han mere economy. By creating apalimpsest of three works to serve as a prcparatory study for a fourth (andindeed many more) Duehamp is approaching the fourth dimensioll in yet

Everything as it moves, now and then, here and there, makes stops. The bird as it fliesstops in one pla<.:e to make its nest, and in another to rest in its flight. A nmn when hegues forth stops when he wills. So the god has stopped. The sun, whi<.:h is so bright andbeautiful, is one pla<.:e where he has stopped, The moon, the stars, the winds, he. hasbeen with. The trees, rhe animals, are all when: he has stopped, and the Indlan thinks

another way-and a very perspICUOUS way from an anthropological perspect­ive, Just recall the title of this work again: The Nelmork IIfSlIJppages. For usanthropologists, those words ring certain bells, because we have Durkheim'sRlemclI/{lI)' Forms of/he Re/iKious Lire, engraved on our memories. And if nO(,then at least we have all read Lcvi-Strauss's Tulemislll (1964) where Durk­heim's words are quoted. Thc original source is :l Dakota Indian, discussingmetaphysics:

:rhe hetter to underline the comparison, Jet us quote without break from the paragraph1I~ LeJ D,·u.l; ~lJllrrC,j' fTh" Tmv SOl/raJ o(Morahty {/!It! ReliKivlI] where Bergson sums uhiS metaphysKs: p

''\ gre.at eurrcnt of creative cnergy gushes forth through matter, to ohtain from itwhat It can. At most points it is stopped; these stops arc transmuted, in our eyes inlOth~ appearal:ees of so many !i\'ing species, i.e., of organisms in which our per~el;tionb~ll1g esse?tta]Jy analytical and syntheti<.:, distinguishes a multitude of clements <.:om~bmmg to fulfil a multimde of functions; but the pro<.:ess of organisation was onlv thestop itself, a simple aet analogolls to the impress of a foot whid1 insmntane~uslvcauses thousands of grains of sand to <.:ontrive to form a pattern. .

The t\l'O accounts agree so exactly that it may seem less risky, after reading them, todaun that Bergson was able to understand what lay behind totem ism because his own:hought, unbeknownst to him, was in sympathy with that of totemic peoples. What isIt, then, that they have in common? It seems that the relationship ft~sults from oneand the same desire to apprehend in a total fashion the two aspects of realitv whichthe philosopher terms (O/lIill1/(1l1s and diJconlillllOus; from the same refusal t~ chooseb~t~veen the two; and from the same enon to see them as complementary perspectivesglvmg on to the same truth. (Levi-Strauss 1964: 98; I(lr refs. see original)

It is surely not difficult to grasp the connection, now, between the subjeet­matter of Duchamp's NetlJ!ork (~fStoppages, and the peculiar manner in whichit has been presented, as a layer-cake of artworks placed Olle oyer the other.The 'Ncc-work', both shows us a I\etwork of Stoppages, and also IS a lIelTl'Ofk0/ j"/oppages, i,e. :l series of 'perchings' at which Duchamp, in his 'flight'becomes visible in the form of an index of his agency, a particular work of an.The NelJPork looks like a 'map' because it is pan of a 'map' of time. But thiscan only be a four-dimensional map. Like Bergson, Duch~lmp downplavs the'merely' visible, or its illusionistic representation. Like Bergson, he dis'trustsour perception 'which is merely an~llytie and synthetic', and seeks instead the'current of creative energy' (i.e. duree, or Heideggcrian 'being') which 'gushesforth through matter'. This is the fourth dimension.

Tn Husserlian terms, there is a most startling analogy between the transpar­ent layering or The Nefmort: o/Stoppages, and the concept of JUfce constitutedOUl of retentions, retentions or relent ions, and so on. The Netmork is a protentiontowards the Large Gla~'J, which is a retention, first, of the original abstractlayout ielf th:lt work (before ils contenr was finalized) and secondly a retention,from this retention, of Utlchamp's Symbolist beginnings, the thirst for tran-

Crmc!miUII: The Erlmded illilld 249

ofthe~e places and sends his prayers there [() reach the place where the god has stoppedand \1'111 help and a blessing. (Durkheim, quoted in L~l'i-Strauss, TOlol/ism: 98)

Levi-Strauss quotes this passage in the course of a discussion of Bergsonwho, he says, propounded remarkably Similar views in CrmtiDI: EVII/utlon andwho was unJouhtetlly famili:·lr with the passage in Durkheim in which the~' arcLjuoteJ. Levi-Strauss goes on to remark: .

F]('.9.s/3. TheYOllng IH"II "",/Girl;1/ Spri1/g b\'l\hr~d DlI~hamp

(H) II)

Crmrlll.lioll: The EXlmded klillJ248

250 Conclu,illn: Tht! t:xunded MimI ClllldIlS;IJI/: The /:"X/(Iulrd tHillil 251

sccndence and release from incest'Uous longing ('initiation') which set him onthe path he subsequently followed. Duchamp allows us to see his 'sinking' pastas a transformable component of his present, retained as something alreadysuperseded in the course of his inlcnerung life. And indeed, there 3fC ampleart-historical reasons for interpreting Th~ Larg~ Glass as a transfonned \"crsion.a distant memory, of The Youllg Man alld Girl in Spr;lIX (sec e.g. Golding 19i3:ch. ).

What I am proposing, therefore, may be called a 'Dakota' model of theartist's Il'lIlire; each artwork, as Duchamp's picml"c so strikingly reveals to us,is a place where agency 'stops' and .lssumes visible form. The Large Glass itselfwas subtitled, by its creator, as a 'delay in glass' (rctard (1/ '-'are), suggestingprecisely this idea (the glass, like ,\ photographic plate) delays the passage ofthe shadows of the fourth dimension, and captures their visible traces. In laterworks the procedure is repeated more explicitly: for instance, in Til m' (1918),which features Irompe~r(pil 'shadows' of Duchamp's prcvious 'ready-made'works (the Birj'c!r-W/rrt! (1913), the Hal-rock (1914), ctc.).

Each Duchamp work, in other words, invites us to adopt a panieular per­spectivc on all Duchamp's works, often by prO\·iding explicit quotations orreferences to past and future \\orks, though also adumbrating retentions andprotentions in a more elliptical fashion. The sum total of the infinitely trans­formable network of internal refercnees (protcntions and retentions) unitingthe f£lIvrt from ,,1/ of these temporal 'perehes'-which we can only, in fact,adopt serially, is the unrepresenr;lblc hut very ((JII((plllo1i:::.ablc and by no means'mystic' fourth dimension. In this way Duchamp triumphantly vindicates him­self as a comic, secularist, psychopomp.

At this point, without having done more than scratch the surface, I curtailfurther discussion ofDuchamp's (L'UVre, though I hope fhat I havc said enoughto encourage any interested readers to pursue the subject further, since the lit­erature and documentation a\'ailahle is unusually extensive and re\'ealing. Mypurpose is only to establish the point that Duchamp's subjecti\'ity, his innerdurrr is conLTctely instantiatctl, as a series of moments, or 'delays' or 'pcrch­ings', in theobjeui\'e [faces of his agency, that is, his artworlsand the texts heproduced to go with them. Here wc ha\"e, in public, accessible, form, the 'con­tinuum ofcontinua of protentional and retentional modifications' described byHusserl for the purposes of elucidating the purely slfbjUlily process of cogni­tion, or consciousness. In other words, as a distributed object, Duchamp'sconsciousness, the very flux of his being ,tS an agent, is not just 'accessible to

us' but has assumed this form. Duchamp has simply lumed inlo this object,and now rattles around the wodd, in innumerable f(wms, as these detachedperson-parts, or idols, or skins, or cherished \'aluables. So we rer~rn to ourstarting-point. Out there is one more step to be taken. Ducha~lp IS (or was)an individual mimI, onc particul:lr person excrcising one partIcular agcnc~.

What about thc art produced, not by individuals O\-er a lifetime, but b~

collecti\'ities over longer periods oftime; what, in other words, can we say about~lIeeti\"C 'traditions'. Can we, so to speak, expand the model that we ha\'eJust constructed of 'the artist's a'1I1:rr' to encompass something wider, SOmc­thing corresponding to a 'cultural tradition' so that we can sec that, too as adistributed obje<:t structurally isomorphous to consciousm.:ss as a tem~ralprocess, or J"rte:

This is the final problem to which I will turn, in concluding the presentwork.

9.6. The Maori Meeting House

Copious data relevant to the quesrion just raised have recently been providedby Roger Neich (1996). He has completed a detailed study of all the extantor photographically documented meeting houses constructed by the Maori ofKorth Island between C.1850 and 1930, which were, in fum, a de\'elopmcnt ofchiefly houses (large structures erectcd with magnificence in mind) of the earl­ier nineteenth century and the pre-European period (Fig 9.6./1 from :\Ieieh,fig. 68). The presem-day Maori 'meeting house' C'Jrne imo its own during thelatter half of the nineteenth century, especially from t870 onwards, a pcriodduring which the Maori found themseh'cs unable to compete with one another(or Europeans) via the traditional warlike melOS. Their comperitivc spiritfocused more and more on the construction of large, elaborately can'ed andpaimed meeting houses, each Maori community trying, so far as lay withinits power, to outdo its neighhours and rivals in this I'CSpcct. The totaliry ofl\-laori meeting houses, therefore, constitutes a particular genre of art produc­tion, over a particular historical phase in the course of Maori history (in J1l,lnyways, a glorious period, which contemporary '\iaoris rightly recall with pride),which can be considered 'coherent' in [he sense we require. All Maori meet­ing houses, that is to say, followed a common 'ground plan', all were dc!.ignedwith a common purpose, namely, to serve as an objectificalion of the weahh,sophistication, tcchniClI skill, and ancestral endowment of the communityresponsible for the construction, and as a ml."ans to ensure that persons notof this community, who might be entertained there, \\ould be consumcd withjealousy and thoroughly intimidated..!\s Nick Thomas has writtcn, apropos of.\laori art:

houses ... were I/O/ 'symlXJI.~' ... bUl vehicles of a collectivity's IXJI\er. Thc~ simulta­neously indexed a group's own vitality ,md ideally or effeelildy discmpowered others.DistiTIl;tions hclwccn function and Illc,lIling, usc and exprcssion, instrumentality andsymbolism obscure what was integraled and processual in these collective prescntationsof tribal efficacy. (Tbomas H)I)5: 1°3)'

, The rcalkr "ould nu. be mis,~Len in lhinLing lh~l 'hcsc mmnlemi h~· Thom"--' pJal·W~" import­~m part in shaping lhe 'iews e~p~ In Ch., "flhe I,resenl "orl.:.

FIG. 9.6/'. 1nc :\bori mceting house~ a rollccthe index ofrommul1lIl pl)IH,r. Soura-: ~l!:a Tau~

e-Waru, Ie Ore Orr _"-Iar:tc, !\lastenoTl. Auckland Public Libr:I.r).

Maori meeting houses may h:lve been the <:ollective production of man): sep­arak artists and builders, working in separate conununitia. at different urnes,each striving to produce somcthing distincti\ e, yet all arc exprcssion~of a c~m­mon historil.':ll trajectory, a common culturJI system, a common ldeologtcaland political purpnse. \Ve arc entitled, therefore, to group them. togcther, asNeich does, since they constituted the 'final common pathway' Jor the phys­ical expression of 'Maoridom' as a (.:Qllecti\'e experience, during rhe relevantperiod. .

Bel.':lllSC thesc collective 'indexes of :lgency' were Itollses-artef.,c.;ts with veryspecial charac.;terisrics all lheir own (c.;f Hugh-Iones :~ml Carstens 1?'J6~­

rhe\ possess features which render them t.'Speclally sUited 1'0 the projectIOnof ~ollccti\'eagency. First of all, houses are 'colJecti\ e' in the simples~ sense.ofall, that people (.:QUect in them, and arc joined IOgcther by them; thiS appliesto any house and is the reason why so many social groups are referred to as'holls·es' (Lcvi~SmlUss; cf. Hugh-Jones and Carstens, op. cil.). Secondly,houses arc complcx artefacts consisting of many separate, standard, parts; theyare thus organized, or 'org'Jnic' entities, unlike, say, a ~wl or a .spear, howe\'erwonderfully wrought. Their organic phm and capacny fo~ dl~assembly an.dreassemblv, remodelling and redecoration allows thcm to obJcctlfy rhe Ol"gall1C

. .. A d fi - II '-and ab()\'c all-thev arcconlll..'C(edness of hlstoncal procl.'Sses. n n.1 ~ '.bodies. The house is a body ftr rhe body. I louses arc ~ies becausc the.y:Ire containcrs which like the body, ha\'e entrances ,md ex liS. Houses are ca\~

itics tilled with li\'in~ contents. Houses are bodies because ~h.ey ha~'e sl.r~ngbones ;llld armoured shdls, becausc they ha I e gaudy, tllcsmefl7.1llg skillS II hlch

beguile and terrify; and because they havc organs of sense and cxpn::s.<;ion_eyes which pecr out through windows and spyholes, voil."(S which reverberatethrough the night. To enter a house is 10 eIHer a mind, a sensibility; espcci_ally if it is such a house as the Maori were accustomcd ro m:lke. Like manytraditional psychologists, thc l'vlaori located mind and intention in the viscera.To enter a house is to enter thc bell}' of the ancestor and to be overwhelmedby the encompassing ance.<;tral presence; 0\ erhead are the ribs of rhe ancestor,in the form of the superbly decoraled rafters which I,:onwrge towards theancesrral hackbone, the ridgc-pok>--the fountainhead of ancestral continuity.On all sides, idols of ancestral beings gaze down hypnotically, entrapping thconlooker in their thought-processes (cf. abO\e, Ch. 7 on idols). The flying sim­ulacra. of the alll.:esrors criss-LToss this interior space with unbelievable rapid­ity and profusion; all merely private, independenl, thollght is overwhelmedand only those cognitions whidl actually emanate from the house, those cog­nitions which are part of the house's vcry structure, are attainable.

Keich pro\'ides a very comprehensive discussion of the cosmological 'sym­bolism' of Maori meeting houses. He shows in great detail how each suchhouse was explicitly conceptualized as the body of the eponymous ancestor ofthe community, who was not so much 'memorialized' in the house which borehis name, as reinstated in this form. The house was nOt a sun'iI-ing trace of theance.<;tor's existence and agency at some other, distant, coordinates, but was thebody which he possessed in thc here and now, and through \\'hich his agenc)was exercised in the immediate present (to describe this as 'symbolic' is obvi­ously a misnomer). At the same timc the house was a multiplicity of connectedbodiL'S, a bodily 'fractal' in Wagner's sense (1l)9.h cf. Sect. 7.I1), since it con­sisted of the bodies of the ancestor's descendanTS, by genealogical succL'Ssion,both living and deceased. The ridge-JXlle objcctifiL'S the gencalogical continu­it y of the chiefly line (notionally by male primogeniture) while the descendingrJ.fters indil.':lle the proliferation ofl.<ldetlines on either side_ These were decor­ated with capti\'ating patterns of the type called J:omhailvha; (Fig. 9.6/2 froml\'eich, fig. 17), which evoked the tendrils and runners of the e\'er-prodllctive,ever prolifer.lting Imll/ara, the sweet potato plam which provided the Maori withtheir staple diet. Body, genealogy, gardens, \\ere all copresent and syncrgicallyaerive. The li\"ing members of the community, gathered in the house, were, so1'0 spe'Jk, only 'furnishings'. Thcy were mobile appurtenances of its solid,enduring structure, into which they "'ould e\'ellfually he absorbed as 'fixtures'.

But the point of this section is nut to discuss the cultural significance of thei\laori meeting house, whieh would be reduml;ln( gi\'en Ihe exccUent work thathas already been published on this subject by '\J"eieh and his colll.-'3.gucs. Thepreceding remarks only reiterate (he thesis, :lrgued in e:lrlier parts of (hiswork, that artef.1<:ts like ]\-laori meeting houses arc not 'symbols' bUI indexesof ageney. In this instance, the agency is collecliYc, ancestral, and esscIHiallypolitical in tone.

253COllrlusioll: Tltr [xtt"dd 11;1/1/Condl/sion: Tl//' Ext~IIJ('l/ !HI/If/252

25... umdusioll: Th~ £Xfrndd JJinJ Conr!usion; 71Ir EX/(Illird ."ind r­-"

1;16. 9-6/~, Kowhai\\'haipallcms; parI ohhe\ttandlh ()I'" porch of Ihelarge huuS/:' of the Ng:ui­Pomu at Wai--<..-Malatini,J::;I~1 C1pc. Smmt: PlaleXIII from I\UguStuSIfamillon, "'lIo,i Art(1&)6).

My aim in imrnducing this material is to explore the theme of 'traditions',To what extent can we study the whole gamut of 1\;1aori meeting houses, dis­tlibuted in Space :tnd time, as a single, coherent, object, distribmed in spaceand time, which, in a certain sense, recapitulates, on the historic.!.1 and colk"(;t­ive scale, the processes of mgnition or consciousness? Fortunately, throughthe very meticulous studies undcrtaken b~' Neich, we can indeed make progre...sin this direction, In order 10 show this, I need do little more than reproducea lable, devised hy Neich himself, which appears on pagc 220 of his book(Fig. 9.613). l'\eich's table 'The Transmission of Selected Figurative P3inringTr.tditions' is organi7..ed in lhe following way. The lefHo-right axis of lhetable corresponds lo the 3xis of hiSlOrical lime (between 1870 and 1930) whilethe top-lO-bonom axis, which is unlabelled, corresponds, implicitly, 10 geo­graphical sp<lce; that is, mccting houses (the large black spotS with numbersand Iettcrs in them) which lie on 3 hori?..ontal axis arc or werc spatially mn­tiguous. The numbers dcnote particular meeting houses in Ncich's comprc­hensivc cat:lloguc of the S3me, and thc letters correspond to 'traditions' ofMaori figurative painting, which began 10 develop and proliferate from 1870onwards-before which time only Maori carving was 'ftgurative' and p;linting(011 meeting houscs) was in the kOUlhaiUlhlll" style and dC\'elopmenrs therefrom.\Ve do not acrualh nL'Cd LO discuss :"Jcich's material in an\" depth, sincc ilis ,'ery detailed ami contextual. 1\11 that I primari[~ wish to do is lO undcrlincami cxplore an obscrval'ion (hat must 11(II'e inevitably occurred 10 the readeralready, th:tt there is a great deal of similar;ty between Neich's di'lgram and the

f"tG·9.6{J. 'I'be .\bori meeting house as an object distributed in spaU' and time. Sou,u: NeichIl}9(,; tahle tilbl 'The Tr.l!1smi~~ionof Scl~ClCd Figurative Painting Traditions', Reprodu<.:ed<.:ounesy of Roger Neich and Auckland Uni"er~;lr Press.

diagram I presenred as Fig. 9.~./I-thc sptlf-and-3rrow diagram 10 which Isought to pro,"ide an abstract model of rhe ':1I'tist's (rm;re'.

Instead ofarrows, Ncich joins the nodal points in his historico-geographicalnctwork by simple Jines; he is not thinking in terms ofprutentions and retcn­tion.<; which, from any given 'now' momem, or from the tempor.tl standpointofany given work ofarl at the moment of its completion, always han a definitedirectionality, I'Owards the past (memory, fI..'C:lpirulation) or to\\3rds the future(project, preliminary sketch). Meanwhile, Neic.;h is thinking, ilS most art histor­i.lIls do, in terms of 'progress'; rrorn the past and towards the future, becausehis whole hook is premised, not unjustifiably, on the idea of the 'developmenr'of a distincri\ c Maori <lrt. So I gues... thar had he used arrows instead of lineson his table, his arrows \\lluld all havc lx"(;1l rightwards-pointing ones. But thisis not rcally logical, in that the concepT of'the transmission of traditions' re.tllyinvolves, primarily, the :ll'list, commissioned to tlcconl1e a house at time 'T-zero''remembering' .In exemplar th.1t he \\itnessed, somewherc else, on some pl'e­,·ious ~uilding, af 'T-minus-one'. The transmission nf a 'tmdition', Ihc reca­pitulatIOn of a model, is the obj(."(;tirk:ltion of memor\' and thus inherent I\"rerrospecti\"e. So the arro\\s ought, mOST logic3lly, puin! 'not from len to l'igh~,but from right til left. But this will nut do either, bec:luse whatever the artistswho made the .I\laori meeting houses had in mind, it certainl)" \\3S not 10

Ctmduswn: Thr E:rltndd Mind ConrhlSlon: TIEr £Xltndrd M/Ild 257

reproduce, tamely, the houses th:l1 had ht:en constructed, by other L'llnllnunities,sometime in the immediate or more distant past. The whole point of thesehouses, as was emphasized earlier, was to bring about the crushing of the(architectural) self-esteem of members of ri\'a1 communities, 10 exah the an­cestors of the house-building community over c\'Crybody else's anceslOrs, byohjecrifying them with superior magnificence and sophistication (which meant,mlt'r alia, incorporating references to pal'tlw-'white'-arr into thc decor, along­side 'traditional' art forms). It is not just that Maori meeting houses inc.:orpor­ared 'innovations' (i.e. non-traditional clements which, later, being imitated,became 'traditions' in themseh'cs), each house was totally an 'innovation' (cr.Wagner 1977) in thai the house was orientated tow::ards the future, the politicaltriumph which would be the {/!/Iicipated outcome of the effort invested in itsconstruction. The building of the house was a collecti\'c, intention::al, ac.:tioll, and'action' is intrinsically future-oricntatcd. The 'agency of the ::ancestor' of whichthe house is an index, is equally furure-oricntated; the ancestor's hodyfhouse isnOl a corpse or a memorial ro the dep::arted. So once :lgain, il seems that weshould m,lke our arrows point from left to right, towards the future. After allthis, it may scem rh:u Neieh has done well to refrain from using arrows <It all, inrhat doing SO resulL" in paradoxes; as 'traditional' artefacL", Maori meeting housesarc undeniably retrospecli\T, as political gestures they are prospcctiyc. Yet howcan meeting houses be both pro~pcctive and rctrospective at the same timc?

But we h,lVe encountered, and (I hope) surn10unLed this difli.culty already.An artefact or event is nenr either traditional or innovatory in any absolutesense, or, as time-philosophers are inclined to put it, SIIb sprc;e aeurt/ital;s. A'traditional' artefact (or e\"enr) is only 'traditional' when \'ie\\-ed from a latter­day perspective, and as a screen, or transparency, through which irs precursorsarc adumbrated. The traditional object is grasped :IS a rctenrlon, a relentiun ofretentions, and so on. Conversely, an 'inno\'atory' object (or event) is inno\'­alOn' only on condition that we situale oursclYes anterior to it in time (i.e. al ;l

mO~lent in time at which it has not yet, or is just about to, come imo exisl­enL'C)-so that we can likewise see it as ::a sen:en through which still laterobjects may be protended, as a protention, protenrion of protel1lions, etc. Thetemporal object constiluted by the totality of the meeting houses displayed onNeieh's diagram consists therefore, nol of a network of temporal relationswhich can be IOtali:t..cd in a single synoptic mapping; but onl} as a 'flle' con­sisting ofa whole series ofsuch mappings corresponding to diffc~e~t l'C.mpo~al(and spatial) points of vantage; each one of ""hich gener:ltes :1 dl~tlnetlve ~lS­tribution of rctenrjonal and protentional relations belween any gl\·en meetmghouse and ils spaliQ-tcmporal neighbours. The logically mandatory .nal~re ofsuch a cOIHinuously shifting pcrspecti\e on tradition and innovation III anhislorical assemhla~e of anefacts means that the process of understanding arthistory is cssenti'ltly akin tu thc processes uf consciollsncss itself, which ismarked, like\\ise, by a continuous perspectival flux.

To express this more concretely, we can interpret :lIly given meeting house,viewed trom the lauer-day perspective, as a 'memory' in objectiried form, ofmeeting houses anterior LO this one. Indeed. in terms of Ihe cogniti\·e proc..-essesof the !\Iaori builders, 'memory' is the facuh~ rL'Spon!.ible for the transmis­sion of the lore, the skills-Ihe 'tradition' in other \\·ords-which the houseembodies. BUI this would not be 'tradition' unless the memory-antecedentsof the house constructed 'from memory' were remembered as having anteced­ents of their own, of which they, in turn, were memories, back into the pas!.Each house embodies not JUSt the memory of its immediate exemplar, hut acumulative series of memories, memories of memories, and so on. That is tosay, it carries with it Ihe whole thickness of duree, and belungs not just toa 'now'-the temporal coordinates of its Jates of construction-hut to anextended temporal field which reaches back into the pasl and which is drawnup into the present again.

Conversely, each meeting house is a project for future houses, a 'sketch'towards a series of as yel unbuilt houses. We arc inclined to see artefacts,especially rather splendid ,ll'l'cfacts like these, as if they embodied the finalintentions of their makers. Bur anyone who has ever had anything [Q do wilhbuilding anything (e\'en an el:tension to a suburban home) \\ ill instamJy recog­nize that this is hardly the case. What gets built is whate\'cr seems the bestpossible compromise in the light of aU the practical difficulties and constraintsentering into the situation; given that the decision to build 'something or other'has all'C',Jdy becn taken. We may think th::at our house-cxtension is "::astly super­ior to our neighbours', but that docs not mean that we \\ould not like to tearthe whole thing down and St:lrt ag;lin, were Ihat a praetic..-able option. We arecert:linly entitled to suppose that the same disparity between ::aspiration andactuality entered into the process of building Maori meeting houses. These, itwill be rec::alled, were erected in an O\-en spirit of competition, in order to

indexicalize the superiority of one community, and its legions of ancestors,over other, neighbouring, communitiL'S.::'IJo meeting house could, in this con­tcxt, be too large, too sophisticated, too expensive and magnificent. But thenineteenth-cenTUry Maori were, as a matter of fact, impo\'crished, oppressed,and diminishing in numbers, while the assistance afforded them by their ::ances­tors was, in the end, finite. The houses thai were built were far, probably, frombeing the houses the .\laon would have liked to build; they were, perhaps,superior 10 the houses which preceded them, they might, arguably, bc superiorto the mecting houses of rival communities (though this would probahly not beconeedcd in pllblic)-hut the) could hardly be superior, or even equal, to whatthey "ere ;mellded to be. There \\ere onl~ 'sketches' or 'prorcntions towards'the llhimate meeling house, \\ hich, for practical reasons, \\ ould alwa} s remainunreali7.able. The mecring house as a 'sketeh' embodied, the promisc, sometime in the futurc, to build the meeting house to bC,tf all meeting houses; andit was this 'threa£' to build, once circumstances became really favourable, the

258 (.'tmrlus;on: T},r Exrr"dd .lfi",1

'ultimatc' meeting house, which was aimed at the neighbours, as much as the'realized' house itselr.

Thus we are enabled 10 see the [otality of Maori meeting houses as a cog­nitj,·e process writ largc, OJ. movement of inner durer as well as a collcction ofexisting objects, and documents appertaining to objL"Cts which time has oblit­eraTcd. The Maori mccting house (in its lotali,..cd form) is an object which weare ahle to trace as a movement of thoughT, a movement of memory reachingdown imo the past and a mm-emenl of aspir.uion, probing towards an unreal­ized, and perhaps unn.'alizable futurity. Through the slUdy ofthesc artefacts,we arc able [0 grasp 'mind' as an external (and eternal) disposition of publicacts of objectification, and simultaneously as the evoh'ing consciousl1t.'Ss ofa collectivity, transcending the individual co!!,il/) ami the coordinates of anyparticular here and now.

BIBLIOGRAPHY

ACKfJl\IA~,J. S. (19;0), The Archirccture of :\lichc1angdo O-brmondsworth: PenguinBooks).

AlX;oCK, C. E. (11)83), Tlfr UirKt Class (Epping: Bowker Publishing House).AI.l.EN, ~,t (19;6), 'Kumari or Virgin Worship in Kathmandu Valley', ComribUllolIS to

/ndia/l Sociol0!J" I olz: Z93-316.."--"Zlt;U, D. (19&», TIlt Skin Ego (New J-Ia\·en: Yale Universit)· Press).1\RcIIEY, G. (1965), The Ar' Forms of PolYl/nia (Auckland Institutc & .1\1useum

Bullctin, no. 4).AR-'Hf..I.'I, R. (1974), An alld Visual I)uaplfull (llerkcle~: Uni\·crsity of California

Press).R.-\BAUI'_",,,, A. (1993), us lJipoullI~sdn dinlx: &soi sur 10 rrlilioll luhili~l/1u Ii l'tpoqut!

de In dhOUt;(rlr (Paris: Maison des Sciences dc I'Homme).IlolIJB, L. (1\187), Red,·mpri'1)c "':lIrolmlers: J'l/re!' Modem Styles ill Ihe lfilldu Tradilion

(Berkcley: UniversiT)· of California Press).RUfIl-lES, R. (19~h), Camua Lucida: R(jfuliullS 011 PI/olography, trans. R. Howard

(I.ondon: Gpe, 19i12).llHESO:'-, G. (1936). NaUII (Stanford, Calif.: Stanford Unin~rsi{) Press).--(U)73), 'Style, Grace and Information in Primitin: Art', In A. Forgc (ed.),

Prill/iii!'/" Art (jnd Socit'ly (Lonuon: Oxford University Press), 235-55.s.....UNO..\U, L. (1972) (ed.), Radical Paspufit't'J ill lht Am 0 Iarmonds\1 orm: Pc1ic-.lO).!lUTftE, J. ([1.116), 'RilOa! and Social Change', Mall. 119: 6o--J4.BtSJA.\II~, W. (11);9), 'On Ihe Mimetic FaeulT)·', R(j/e(/ions, 00. P. Oemef"Z, trans. E.

jephcotl (New York: llarcourt llra!.:c JOI ano"ich), 333.BERGER, J. (1972), Wuys ol."'ecillg (I.ondon: BBCjPenguin).RF.ST, E. (IQOI}), 'Maori Forest Lore', Tral/saC/iollS 01 !h( IVrJl' Zealand bwilUft, 52:

433--81.BloVAN, E. (1940), Holy 11/fQKN (London: George Allcn & Cnwin).BUlK IF.. J. (1914), 'Imabre5 and Idols (Eb'YPti;lIl)', in J. Haslinb"S (cd.), El/Q'dopat!dia

ofRdigitll/ amI Efltirs (r:.<linhurgh: T. & T. Clark), lUI. ;, 131-3.BLOCH, M. ([991), Prey inlo }-filiI/a (CambriJge: C1mhridge University Press).RcHS, F. (1927). Primilii'( Ar! (rcpr. 1955, New YorL: Dovcr llooks).ROI!RDIEU. P. (1968), 'Outline ofa Sociological Theory of ,'\rl Pert.'Cplion'. lutrmlJtiOl/al

Social Scim(( Juufllal. 2014: s8Q---6 12.-- (1977), Ollllil/( 0/1/ TIt(ory o/PraCli(( (Camhridge: Cambridge Unil'crsity Press).-- (1I)8.j.), DiJIII/{/i,lll: A .'Ii/rial Criliq/le 0/J/{(~~emellts ~r TIl.IIe, trans. Richard l\'ice

(London: Ruutledge & Keg-an Paul).BOYER. P. (1994), TII( NlI/um/lltH of Rr/~~iou5 "'(as: A Cognitil"t: T!I(ory oj" Rrligio1/

(Berkeley: 1jnh·crsil~ of Califomia Press).(1996), '\\'hal '\1:ll.cs Anthropomorphism i'\alUral: IlIIuiti\-c Gnwlo\!:\ and

Cultural Representations" JO/lrnal uflllf" Roy/I! Allthropological/lls/illlte, ~II :'S3-(}7.CA.\ULLl,1'\'t (1989), "11/(' Gotll/{ Idol (Cunhridge: Cambridge Unh'crsit) Press).C\IlliY. S. (1985), C(II/api/wl Challge in Child/wot! (Cambridge. :l,.bss_: ]\11'1' Press).

26(l

CARPENTER, E_, :lnd SaIl.ST£R, C. (t986), .Matrrialsfor lire Study ofSocial Symbolism illAI/riel/! and Tribal Art: A Ruord IIfTradition alld COII/ill//il)', based on the rcsearchc.~

and wrilings of Carl Schuster, 00. and written by Edmund Carpenter assistL'(1 byLorraine Spiess (New Yorl.:: Rock Foundation). .

CoOTE, J. (1992), '11arvc1s of Everyday Vision: The Anthropology of i\csthc:t1cs andIhc C.anle Kc:cping ~iloICS" in J. Coole and A. Shehan (cds.), Anlhropology, Art and

Aeslhel;cs (Oxford: Clarendon Press), 2..j.5--75·__ (uJ9fi), 'Aesthetics is '1 Cross-Cultural C1tegory: For thc !\'lorion', in T. Ingold

(00.), Key Drbalrs in Allthmpology (London and New York: Routledge), 266-]5·OALLAPICCOLA, A. 1.. (1989) (cd.), Shas/rie Tradi/ions ill 1I,diall Arts (Smttgarr: Steiner).DANTO, A. C. (196..j.), 'The Artworld', JOllrnal of Philosoph)', 61: 57 1- 8 ..j..__ (198 I), ]7,e Transfiguration ofthe COII/mol/piarr: A Pllilosophy ofArl (Cambridge,

l'vbss: Harvard Cni\'crsitv Press).D'A2f.vroo, W. L. (1973), ;:\'lask Makers and Myth in WesterJI Liberia', in A. Forge

(00.), PrimiliL't Art alld Sociel)' (London: Oxford University Press), 126 68.__( 1973) (ed.), T/lr Tradiliollol Ar/isl in Afriroll Soarlirs (Bloomington and London:

International Affairs Cemer).DEACON, A. B. (1934), 'Geometrical Dr'lwings from Malel.:ula .lnd Othcr Islands of the

I'\c,:w Hebrides', JUllrnal oflhe Royal Anthropoloxicallnsti/lfu, 6..j.: 121)-'75·

DF~""'"NE1"T, D. (11)79), Braim/orms (London: Harvestcr).DEN:-.~:rr, R. E. (1906), At Ihe back oflhe black man's lIIilld: ur nOies on Ille Kill!h Office

in West Afn·ca (reprinted 1968, London: Haarlem).DOR1)ILLQ:'<", R. I. (1931) Grawnli/ire el diClimlllaire de la !llI/gue des 11u Marquises (Paris:

Tnstilut d'Ethnologie).DOL"GI.AS, ,\1. (1966), Pun·l)' aJld Dul/ger (1 .ondon: Roulledge & Keb":ln Paul). .DU\ll:ZIL, G. (19&), Call/i/ll/s: A Slud)' of IlIdo-EllrOPUIII Religion as Romllll His/or}',

trans. A. Arona" iL'Z and J. Bryson (Berkeley: uni,-ersity of California Press). .I::<':K, D. (1985), f)aNrm: Secillg fI,e Divine Iwage i" India (Chambersburg, I'a.: Amma

Books).£Co, UMK£RTO (1976), A Theory ofSemantirs (Bloomington: Indiana U?iveTsit)' Press).__ (198..j.), Semiotics 011I11111' Pltilwlphy ofLungl/age (London: !\'1acmIU~.n). .ELLL'i, W. (1831), PolYlltsion Reullrrhes, ..j. \"ols. (2nd oon.; London: hsher, Son &

Jackson)_ . _. .E'iCH\"\.'N, A., el 01. (1978), Cull ofJogal/nufll and tlte ReglOllal Trallt/lf//I of Omsil

(Delhi: Manohat Press).FARIS, J. (lInl), Nuba Persol/ul Art (London: Ouckwonh). . . ..Ff..RNA......m:z, J. (1973), 'The Exposition and Imposition ofOrder; .~rt.lstlc ~xprcsslon ~n

Fang Culmre'. in W. T.. d'Azen~do (00.), The Tradilional ArllSl In AfrUIIIl Sooe/les(Bloomington and London: Indiana University Pre.-.s), 19..j.-2-:0.

F".....ol.Ay, J. (UJ75), 'Husserl's .'\nal)sis of the Tillle Inner ConscIousness', ,\fOf//S/ 5911:

J-2o. .FODOR, JURY, A. (19l;l..j.), 'Fodor, Jerry A.', in S. Guuenplan (cd.), A COll/pall/lI1I /0 Ih~

PhllosophV ofMind (Oxford: Dlaekl\ell), 292 300.

Fo~m, A. -(1966), 'Art and Environmcl1\ in the Scpik', Procttdi".~s of Ih~ RoyalAlIthropologicallllS/;I/IU ofGTtUI Brililill alld Irelalld for 1 ¢i5: ~3-J1..

__ (197J) (00.), Pr/milil't Art IIlId SOOj'f)' (I.ondon: Oxford Unn·ersll) Press).

Bihlif)gruphy 261

FRAZER,j. (1980), Tilt: Golde" Bough: A Sillily ill Magic (Jnd RdigilJ/I (London: Macmillan),FREEDBUG, D. (1989), The Prlll'rr oI/magtS (Chil.."3go: Universir) of Chi,,;ago Press).GAO,o\.\\[R, II. G. (1979), Truth and Mrthod, trans. and ed. G. Barden and J. Cumming

(London: Shecd & Ward).G[U~ A. (1978), 'The Gods al Play: Venigo arlll Possession in Muria Religion', Mall_

"515: 219-48.--(199zu), TIt/' Al/throjW/olD' of TiN/i.' (Oxford: lkrg).--(I99zh), 'The Technolog:. ofEnchantmenr and the Enchantmenl ofTcchnology',

in J. Coole and A. Shelton (cds.), AIIIIII'Opo!OKY, Art (/lid Aesthetics (Oxford:Clarendon Press), {0-67.

--(1993), Wruppil/g in Imagts: Tal/oo;n! i/1 PolY/in/a (Oxford: Clarendon Press)."-- (1994) n.d., TaU'ooing in India. Paper prescmeu to the World Archaoologil:;ll

Congrcs..~, Delhi, 1m.-- (HJ95), 'On Coole's "~\'fands of Everyday Vision''', in J. F. Weiner (cd.), Too

IHIIIIY Meunings: A Crilique oIlhr 1JlliliropoiolfY of AeSlhelin, special issue, SocialAnalysis, no. ]8, pp. 18--)l.

-- (1C)t)6), 'Vogel's Ne,: Traps as Artworks and Artworks as Traps', }lJIlfIlIlluI

AlllterialCullurr, If I: 15-]8.-- (forthcoming), Ti,e Art of Alltltmpology: Essa)'s alld Diagrllms bV Alfred Cel!

(London: Athione)_GuJ.., S. (1992), The Chtlflll in Alliria Suriely (Reading: Harwood Academic

Publishers).Gmso:,\, J. J. (1986), The Ecolo.'I,iral Approach 10 Visual Percrplwn (Boston: Houghton

Mifflin).GOLDI:'\G, J. (1973), Mural DIIChll1llp: The Bride S/ripp<'I1 BuTt by her Buehr/u" n·ell

(London: AlIcn Lane).GD:\lllRICH, E. (19R..j.), Thl' Sense ofOrdl'r (London: Phaidon).GmlflRICH, R. (1966), 'The Consecration of a Buddhisl Tmagc:', Journal of AsitJn

Studies, Zll/I: 2.l-:J(J.GOODE-"IOUGH, W. H. (1l156), 'Componcmial An'llysis and lhe Study of Meaning',

Language, 32: '95-216.GOOOi\IAN, N. (1976), Lungllllgtf ofArt (Indianapolis: I Tal:kcn).GVOIRIE, S. (1993), Fam in the Clouds (Ncw York; Oxford Univcrsity Press).IIAl'JOY, W. (UJ22), 'Tauooing in the Marquesas', Bishop .MU~lIlIl Bullrlin, I

(Honolulu).IIANSON, F. A. (1983), 'When the J\'!ap is the TerrilOry: Art in Maori Culture', in

D. K. Washburn (ed.), Stn/{/UTt alld Cognitioll ill Art (Cambridge: CambridgeUnin~rsity Press), 7..j. 89.

HARIUSON, S. (1985), 'ConL'eplS ofthc Person in Ayatip Religious Thought', .Mun, 20/1:

1'5-30 .I-1f.-"':RY, T. (H}28), 'Ancient Tahill', Bishop IHuSt'lml Blllldin,..j.8 (Honolulu).l-hRN, \'IIJO (ll}Oo), TIle On!lI/s ofArl (repr. 1971, Ne\\ Yorl.:: Benjamin Bloom).HOFSTAIJTF.R, D .• and D~;N:'<"F.l·r. D. (198z), The .Mimi's I: PI/lltl/sics I/Iltl Reflerliolls 11I1

Sdflllid SOIlI (Harmondsll"orlh: Penguin).HOLLIS••\1. (1970), 'RC:.lson and Ritual', in B. R_ Wilson (c:d.), R(I1iollality (London:

Basil lllacl.:\\"cll), 211-39_

262 Bihlioxrapny Bihlio!raphy 263

1-loLT, E. G. (19;7), A Documentary lJisl0'Y of Art, 3 vols. (princeton: PrincelOnUni\"crsiry l'ress).

HUGH-JONES S., and CAKSTENS, J. (1996), /Ihoul Ihe: HOllse (Cambridgc: CambridgeUnivcrsiry I'n:ss).

IRSE.'X, II. (1966), Pur Cyllf, (rans. P. WaIlS (Hmnondsworlh: Penguin).KAEPPUlt, A. L. (19j8), Artifitial Cutlgsitia: ReinK an Exposition ofNative MO/lufar­

lum Co/luud on the Tlrru Purifit Joumqs ofCopla;"Jama Coof at tire Brmire PI/an,Bishop .rHusellm (llonolulu: Bishop \1uscum l'ress).

KORN, S. M-. (IIJiR), 'The Formal Analysis of Visu:11 Systcms as Exemplified by aStudy of Abclam (Papu;l New Guinea) Painting', in M. Grccnhalgh :lnd V. MCgilw(cds.), Art in Sotiety: Studies in Slyle. Cllltllrt and Atslhelin (I.ondon: Duckworth),

161~3·

KR..UlRlSCH, S. (1946), Tire Hindu Templt,2 \·ols. (repr. 1976, Delhi: ~\orilal Banarsidass).KRL~, E., and KUIlTI., O. (1971), UKI:Nd, Mylh and Magif ill tile {magt ufllre ArtISI: II

His/fm-cal b:periment (New Hal'en and London: Yalc University Prl:Ss).KOCIILf.II., S. (11)8;), 'Malangan: Art and ,"temory in a Melanesian Society', Afllll, 22/2:

238-55.--(1988), 'Mabng-.l.Il: Objccrs, Sacrifice and the Produl1:ion of Memory', A",~ritan

Ethno/uxist, 15/4: 625-37.--(11)/)2), 'i\·laking Skins: l\'lal:mgan and the Idiom of Kinship in ~onhCIn l'ew

Ireland', in). Coote and A. Shelton (cds.), A",hrQPology, Arl amI Aesthttirs (Oxford:Clarcndon Prcss), 94-112.

1....."'GSDOIWF, G_ 1-1. VON (1813-1"), Voyagrs and Trurtls ill V/,rialls Paris uftht World18aJ-I80i,2 vols. (London: Colburn).

LA\"ONots, H. (1996) 'A Polynesia.n Game of Swings' ,Journal oft"t PU!Yllesion SMitty,105(2),201-16.

L.-I..YARll, ). (193(), ';\'Laze-Danccs and thc Ritual of the Labyrinth in i\-lalckula',Fufklore,4T 12J-jO.

--(193j), 'Labyrinth Ritual in South India', Polk/rm,4l): 115-82.LEAD-I,)., and LLACH. E. (1983) (cds.), Tilt Kula: New PtrspettiVfs 011 J11assim E.fthlmge

(Cambridge: C::ambridgc Lin.iwcsily Press).I j:v-.-Sl1l.AUSS, C. (11)(J2), lA Prnsit SO/lt'aJr (Paris: Pion).--(1<,(63), Strurtural AlllhropfJlogy. frans. C. Jaoobson and B. G. Sc.:hocpf (1\ew York

and London: Ihsic Books).--(1904), 7"0/(/Ilis/II, trans. R_ Needham (London: Merlin Press).__ (19/")(), The fltll/enlllry Stnlr!llrrs ufKillShip. trans.). II. Uellt,). R. \"on Sturmer,

and Rodney Needham (Roston: Beacon Press).-- (1970), TnI' Rull' IJIld tht Cookrd (London: Cape).-- (1972), Tht Suvagr JUilld (London: Weidenfeld & :'\Iicolson).LUc.Rt:-nus (1952 cdn.), De Rerum Nalllra, IrallS. H. A. J. '\1unro (Gre:n Books of the

Western World; Chicago: Encyclopaedia Britannica Inc.).j'V(AJ.lNQWSKI, URO.'1ISL... W(1922), ArgMulIls {/Ith~ Wesftm Paojir: An AUOIIIU ofNlIlh'/!

El1lrrprisr alld Adt;ttlf/lrt i" the Arrhipda.v,os of .Iltlonesia/l Ntw Guinta (London:Routledge).

__ (1935), Coral Cardms u"d tlltlr Magir: A Study oflht ,\ftthads of Tilli',g Iht Soil(wd ufAgrirulilirul Rites in tile Trohrilmd Ishmds, 2 \ols. (LoIlllun: Allen & Unwin).

i\'hRR. D. (19~h), Visioll: A Compl/lalional blusligatioll into Ihl! fJuman Rrprestlliali/JII(Iml Pmce5sil/g of Visuulll/fimnritiof{ (San Francisco: Freeman).

MAVSS, .1\1. (llJ02),A Cell.'rtll Therl/Y oIJ1'IIIf(if, tran~. R. Brain (repr. 1972, New York:l\oTlon).

-- (11)5..), TIlt Gift (London: Cohcn & Wen).MUllJt, D. (l98j), Maltrlal Cullure flmi -'lass CO/lSUmplion (Oxford: Basil Dlackwell).,\IOOKF-RJt:F., A., :and KJIA"'N-\, ~t. (197i), Th/! ]illltrir Way (London: Thames &

Hodson).;\tORPIIY, H. (I(J9 J). AllcfS{ml COIIIII:f/i'Jl/s: ./Irl IllId UII Aborigina! 5y.\·/cl/1 /)IKnoll'{ctflfc

(Chil:ago: Chicago Univcrsity Press).--(1994), 'Thc Anlhropolog) of An', in T. Ingold (cd.), Compallion Ellfydoptdia

ofAntltropolog (London and ~ew York: Routledge), 648-85.--(1996). 'Aesthelics is aCross-Cultural C...:llcplrl", in T. Ingold (cd.), Kty Dl!baus

in A",hmpulov' (London and New York: Routledge), 2;j-60.--(1992), 'From Dull to BrillianT: The ACSLhetics of Spiritual Power amon~t the

Yolngu', in). Coote and A. Shehon (eds.), Allthropol/)gy, Art (11111 Amlmifs (Oxford:Oarendon Prcs.~).

\ll!NN, l'\. (19jJ), IValbin· lamotruphy. Graphit Repnsrntation and Cultura! Symbol;smill a Central Australia" Society (hhaca, 1'11': Cornell Universily Press).

--(1977), 'The Spatiolemporal Transformafion of G:awa Canocs'.Jolirnf/1 dt Sotiet;deJ Orftll1istcs, J9-5).

--(1986), The Fallli' IIIGIITI'II: A .~)mb{}li( S/JII~)' IIf Villue Trllllsji"mlllill/l ill II Mussilll(P.N.G.) Suritty (Gmbridge: Camhridgc Uni\'ersity Press).

NDQ!., R. (1996), Painttd lIist{)tlrs (Auckland: Auckbnd lJnilcrsiry Press).PARIty, j. (1985), 'Tradition and the Tcchnology of the Illlellecl', in J. ~'cring (ed.),

ReasDII und tHura!it)' (AS.'\ \lonographs, 24; Lundon and Ncw York: Tavistock),200-2;.

PRICE, S. (1l)89), PrimiliVl' /Irl ill Ch:iliutl PlllffS (Chic;lgo: Chic:lgo Univcrsity Press).-- and PRJo:. R. (IQSo), AFo-Amfriea" Arts of Ihe Surinam Raill Fortst (Berkelcy:

Univecsit}· of Glifornia Press).S.\lILl'\S, M. (1974), StOHt Agr Etonamits (London: T a\·istock).SO-l.ll:R, F. (1986), Duper into PicturtS (Cambridge: Cambridge Uni\'ersit)" Press).SINHA, j. (11)34), II/diun Ps)'ehoIIJ!{)' ofPerup/ioll (London: Rootlcdge & Kcgan Paul).SPERREIl., D. (1985), 'Anthropology and Psychology: Towards illl Epidemiology of

Reprcsentations'. MaJJ, 20: 73~98.

STEINE..";, Karl I'on den (11)2;), Die IHllrqucsallrrlllJd ihrt Kllnst. 3 \"Ols. (Herlin: DieuichReimer (Ernst Voh.sem».

--( (988), 1/ull den S'eillrll 's '''orql/('son Myths, cd. M. Langridge and rrans. j. Terrell(Canberra: Targel (kC',1niafJournal ofPacifie Srudics).

S·l"f.I;o.iER, C. B. (1994), Ajdeoll Art il/ Tramil (Camoridgc and Ncw York: CambridgeUniversity Prcss).

STEI:-;~IA;o.iN, R. (lg89), 'Kolam: form, Technique and AppliC',1tion of a ChangingRitual FoiL: Art nfTamil "ado', in A. L. Dallapiceob (cd.), Slrastrit Troditiul/s illIndial/ Arls (Stungan: Steiner), -l-i5 9 T.

STR.nltER'<1.:\. (19il), Tlrt Ropeof.lloKo (Cambridge: Gmbridge Uni\ersil) Press).STR....n IER~, .\1. (1988), TIle Ge-nder oflhe- Gift (Berkeley: t;nl\erSil) of California Press).

21>4 Bibli(J~raphy

INDEX

TA.\1Il1AII, S. J. (1985), (,'u!l/lf(, Tll/JugJI/ and So(ia! Action (Clmhridge, j\tas.".: IlarvardUni\'crsiry Press).

T"USS1G, ,\ I. (1993), AI;/1Ils;s and A/ltnl)' (I.ondon & New York; Routledge).THOMAS, :"J. (H)t)O), ,Hurqlll.'Sfln SUfitflrs: JIIt'quallly (/1/(/ Po{iu((/! Tra/lsforlllatiQII IfI

F.as/all PolYl1aiu (Oxford: Clarendon Uni\crsity Pm;s),THO,\IAS. "N. (1I)!)5), 'Kiss Ihe Uaby GIMJdb)c: Kowllfliu>/lUi and Acsthclics in AOlearoa

New Zealand', Critiwl ll/qljiry 22 (AUlumn): QO-121.--(H)91 l, E/llllllglrd ObjrC/J (Cambridge, '\·1a."s.: Harvard L;ni\ersit~ Press),THO\lPSON, 0'ARCY (1()6!), 0/1 CrOll'l/1 {Jlld Foml (Cnmbridgc: Cambridge L"nivcrsity

PrClt'».

T~IOMI'SO", R. F. (197.\). 'Yoruba Artistic Criticism', in W. L. d'AzcH:do (etl.), TiltTradil;ollnJ ArtiSl in Afriall/ SQ(irlirs (Bloomington: Indiana univcrsity Press).I g-(' J.

TREGf.AR, E. (,891), The Maori-Polynesiall COlf/pafOllty DirtiolUuy (Wellington: L}on& Blair).

TURNER, V. (1968), Drams ojAff/iflioll (Oxford: Oxford UJliH~rsity Prc~~)_

Tn.olt, J. (1875), Primith'/: Cllltllre (London: .\'Iurray).WAGNER, R. (1977), He Im.Yllfio" ofCaltllrt (Englewood Clif(~, '1J: Prentice I-Iall).--(H)91), 'The Fractal Person', in'\l. Str:llhem and '\'1. GOOelier (eds.), /Jig AIr"

01/11 Creal Men: PefSOJliji((II;OIlS of Pal1'ff ill Mela'u'sia (Cambridge: C.ambridge

Univcrsity Press). 15(}-73.WA.~HBUJ(X, D. K. (198J) (00.), Str/If/are 011I1 Cogllition in Ar/ (Cambridge: Cambridge

Uni'ersit} Pres~).

-- and CROWE., D. W. (1992), S)'JI/"'f'lrin ofGallltrr (2nd cdn.; Seattle and London:

Univcrsity of Washington Press).WIXOI, P. G. (1958), TIlt Idea oja Sorial S(irJl(f lind iii Relalion to PIII"Iosopl,y: Studits

ill Philosophiral Psyrholof,)' (London and Henley: Routledgc & Ke<"oan Paul).WrITKOWEIt, R. (1978), Idea I/Ild Image: SIUllta ill Ihe IIt//ion Rtlloissf/IICI: (London:

Thames & Hudson).WOU..II£I:\I, R. (1987), 'PiclOri:l1 Sl}"lc: Two Vicws'. in Berel Lang (cd.), Till! Conapt of

Style (Ithaca, NY: Cornell unin:rsity Press), 183-202.

A·a '37-.P. ~18I\bcbm "'-2~Wuetion:

"f ~gt:nC}", pcrs"nh'HI<1. and inJe~ 222-.1,227,130 I

:asoogniliu, process 13-,6of iho:: lks!in} of arl ..."Orb Z-4Index as csscnrial for 36..f,he Protolype ~5, '0"

AJorm>, T. T. 99atSlhcria;

IUld CO/lCrpl of~-k: '36--'1, '39critique "f 2-", 62, M, 7", 11l-2, (,I", 95, ()7.

'55, 1;6culn>n:/pcriotJ specitic I----Z. Ib,97. 139a.s ~ form of a~fK} 3,.66HaMOn Ihcury 137-fIoas incorpordlcd within thc<,r}' 74-5. 1;.1-"Ka1llian 8~

as 5",,"'1atorial respun«e 5,8,--2s~·Ic. as a form ofX!>"'fhclie personhood

'57·unh·cn<.11 ~c'lhelics' of pallerns ,60-1

a!linal rdations 213,226agen~-:

:><:stllclic2~· 31.66"farti,t 23,28-30. 33..H. 38-9. 51, 5S-I:I,

68-'72

~'defining an 7,66,96,1.:;3of human beings ill relation ro God II.f.' 16.

118, '29of me inuel ~s '<:mlral' bUI not 'mUSI

important' ;;1\within it\Jular,wor~sin()n-li,·ing things

17-'9.20 I. 28 J~...6, 5'. M--'18.l)6, 1:13.",and inlentionalil~ ,6-17. 19.20 3,67 8,

'26-.lJilllernali,rjeXlcrnalis, thcorit:.~ of '26-33,

1.fO 1.222. 225-6inmhing cugni,i,e process of ,.Wuelion

13 ,6.2". J6. 83'pa..,i,c' ag>:nc)' HoO 7, 110 '1.12'),135ufpatron -1 .....7-8primar) and seeond2rj .g...n~~ '7.20-'.35.

368.51prole<:li,-c agene) 83 1)0

religious 96 9rCl"odllc..,d/pcrpetuOlcd 22;; 8of speclalor 33 4. "f, 2'<uhjecti,il)/obl~....ti,-it) of 56-7,67 8.78

ali lransili,,,/rc!a1iunaJ CUncepl 22,2", 51, 58.62.6"-5.68,78. 116-17.123, J3..--'1

,-\lIen, ~l. 67. '50-3apotrup,ic art 8J-90Il1Iilllll' 187

aniconic rcprcscntllion J6.97-9> 110, 131-2animation!JnimaC)·:

of Christian church 142 Jof idols 1:11-6, lJ6-7ofin>agc; 118-2'personification "f phnrs -II..·irhin pmems ;6-80

anmropol"gy:"hrt I '[,83,96-7, '53, 157, '51), ,60-1and Ih" human life '"yele 10-1 Iprototypical~ of 9-'0in relation IU Sociolo~' 8. '2sulij,..,t nllmer of .1-5,8-1" ,6-, 7

anthropomorphism 59.1)8-9.121-(" IJI---2,

135!\ro;icu, D. q7.\rchc)', G, Ifloan:

abslracT contraslcd with rcpr"...elltatiunal

'.10( 11>1: bod) ,(,8--90. '91, IQ-l

and c~>gnilh-" TC!iiSLancr 85-6,95rlcf..·n<i.-c 8J 90,20<) 12

anuexcrciscs"fcomparison '59. ,6[, ,6Jhislori"ns 2,7.150fDrnl,,1 anal,-,.;" 16.1-5. ,(,6--'1. 170as a fi,m, of"cring 6]as ~ form of WeSlcrn iJ()I~tr}' 6:1 -to 97and id"lJlry 96 '). '0<) 15"nd magic 71-2.99-'0',103-":IS nupping cultul"(" 159 62.215-"20IlJrra[i'-e 21 [-qn"ll-rcpre,enlati'OI1,I/de~""'ati,C 73-1l6,

,60pcrfOmLalKt' art (r,psyc~ical/l'O~niti,-"$tllima of 157, '59.

215 20

in relib';OUS CllnTe.", 9ti-'), ,06-26. 150'tr:ltlitioll'/innmalion '35, '58. 233-4. 23'.

2$-l.256Weslem in 1"("1.lioo 10 non--\\·eslem 3,5.8.

30-' ..H. -11-2, 59. 67. 7.•. 82. 97.158-9,,6I1n.. '9',20'.2']

women'> 7.1.84 55U IIlso "e<;lheti~. <lyle; "n,hropt)l~: an

obj,'Ct; 'nde< •

266 Index ludt:r 267

ulist:as ~cstheric agent 66as ·bali,....n,.,,' S8-ljccntr.dity of ~2

a.~ crtali,-", asen! 2]•.03. lS, 38-9. 5', fot}:as 'c;unning ago:m' -U-Se:<ercisinl p",,'er o,er spe<nlon; 39. 60), 71-2as 'hero' 39> 72:as idolalorjiconocl.tS! 62 5Il'lIrreof 4(, 7, 15,l,'!.!·z 51'S par;,i,'c agem 28-:>0. Ji•.19, 4(,possibility of sc,-cr,d of one single inde~ 64-5as §imuhaneously Agent ~",r Palienr .19-4°,

45-'7. F. 1.19WC'>I.em 30, .n. 35•• 5]-8women as i J, 14 "IU IIlJ~ under indi<-idual names for sp.:cilic

Wt'Stern utisl~

ut ohjt"CTs/"orh (termed 'index' bcfan p. 153)::ltta<;hing I"'TlIOl1S [0 thinK~ 74-5, 80 2as ,,-",,[acl,,:'" ~rsollS 7, .2, 1;-19. 62 6,153.

21.0, ,ns-f1, 2:8. 250 T

as 'bodks't-dins' 6,. ':;;-53. lOU n.• U5-8as CUhUT21 objo:ct5 153 -f, 159-62. "'5-wdefined 5.7"-9, 12:l' the ro,d ofsoci~1 agency 18-1!!. lJ, 28-J2,

e.G, II], ns-6as magj~-al cmllions 6<)-jl

obj~...1if~-ing Il1cmury !2J-8. 1j7objc:l.'llf)'in~social ~btiolhhips 62, .Uj 6,

219 30as pnmal')' Ag..-nts 46. 14Q-50as =ndu~ agenlS 1;,20,,~igns $-7, ~fJ, 165, l~fJ

statllS within anthropology of ~rt 12. ]5-8, 150tenncd 'the index' 12-[0.2J-4as tempol"3.l objects 2J5 jl

Amhcim, R. 43Al')'a Samaj 1.24.~tfJ·31

D~badzan, A. 106-1;1, lID, II.', lJ7 n.Dahb. l~ t 16l.Iax:mdall, L. 2Ihma; Collectiun 47IbltsOO. G. H.I34Heanie.}. 101Iknjamin, W. 31. QQ-loolkrgwl1, l-l.-L. 'l,p, 249)Jcrg"r, J. ~

Ikrnini H-4lksL. E. 106&..,.n. F~ 124hirtb:

I\brqul'$an SWIM: image of dcil~ p;i'-mg bIrth

'"mylhufKae 2IJ-I"

biov.ophy:within anlbropolog~' 10 "shared between ~rt "ork and pt:r.;,ll1 05,81

Illake, William 39Illoch, ,\1. ('2llolls. F. 4. 156IJourdM:u.I'. H.42, 127lll;>yer, P. 1.1.. In, 126. 12;, 131Buddhisl:

land.'iC:lpoe anists .10.)1cons""ralion rill." in Sri Lanka '¥i-51

Ilraqu". G. 2JJ

Gunille. "'I. '41-2Ca,weno 42capt;'-"Iion 68,2,80C-a~r. S. 1:1.[C,arpenter, J::. and Schusler. C. 7Jll1ste spcdalisls 145-7, q8-9, 15!lllllsalit~:

and ag"ncy 16-17.19-21.25doma.ins of wilhin an wm-k -tJ -I, ;6-;as milieu )1-8,131-2ami mimesis 100 I

and perttpliun JJreprl'sent~tiunof ")-.:;.nd sorttr}' [04, 108

l'a"e paiming5 33C"elric knol"urk 8..CU.:lnllC:. P. I~. lJ). 2#

Qurdtill. \\, 49.:bssification of ~'Ie 156-8cognition:

.nd Ihc ~hdllcrioll of ~genl:' 1.. ,.16~nd anlhropomorphism 121 2blocbg~ in relalion to art work 71----2, 85 6,

88-",mil o.tneq)t of tlllr;, 2J6Kub as a farm of 232and mateS 882~sociolit} 17,75

Congu 59l~)Ilsecr~tion rituals 14.l-53C..nsu.ble, J. !33C..uure.j.l,jCf'C2ti,-il~·. Westem and non-Western ideas 30.

33Cubist 25, 242-3cultu"':

as an.logue of person Ij8-92nd ~Illhropolog} of~rt 2--1, 10. 11.33. t53,

[55and ,;sual st)k 155-62. ZJ5----20

eu,itt, lbf()fl G. L. 166

dane<.' 93""'5D,li. S, 55-8

mnw, A. C. j

Ja"I"ln 116----21,129,136, I-Wf)'Aln'edo, IV. L. 46Dcacon, A, ll. ')0 4de-alh:

and:m forms as p;lths 10 lhe undcnmrld

"'~Butldhist ima~ as reliqll:uy qS 9oflhe image 04, 145-6, 22J--Iaod ~Mang-ao rill-'S 223-~

im-oh-ing- nail fctishes 61-2.nd the social [>Crsun~ 195

dCl.~)l1IliI·e atl 73-95. 160-1DcnllC:t1. n. ij. IJO-I[knllC:ll. R. E. 60-1design 73diffusiunism 87 8,94 5distributed:

objrdS 22[-5~

personhood 21.103,1"4.2.10dolls, playing ,,-ilh iIS cult bc-hoa"iuur 18.134Dou~bs. M. !Izd",,,;ng 45.46

as a form nfdance 94 5and artistic representation 190 I

dre.m imagery 58dualisr thuughr 127Duchanl..,M. JO-I,22I), 242-51Duc:cio 47Dume-lit. G. 9S'1fT« 2J6, 243. lMl-jo. 258Dllrkhe'nl, E. 248-9

Cllrplugs 211 '-tEck, D_ 116-1;, u6F..co.li. 1;4. 15F:!)'ptian an/idols 2-1.87. IJ3-... 1-t6 n.Ellis, W. III

o:motions ob;c-cufied in a" worL. 6,31t-nchanUllenr .11. J9. 58, 68-'72, 74, So, 83.

8;;-6Epieurc~n theory 10..-6,108, 11;,223F.s<;hma'1l1.1\. 144tlllll (:\larqUCl<Jn molif) t71 ll4e"olutiunism 611-9C),dtan~

in coli of jag:llUl2lh -I5~0

ofdi,-inee.,u,'itt Ill-I], 146, [~6n.

and origin.tion 58process within ~orccry [06 8see also ""Ia

cxm'iae III [5. ,,6, I~6n.

eve'. 'discriminal~" 1

,nd d,., establishment of inter!lllbjecti,-il1"'",,6 Zl

'e,'il" ,16

p;liming of on Buddha image 1-19as S~'mb()l of imcriurill" I.H-6. l~fJ-7

'Ihiru' of K,mlllFt 152

fan handles 202 -Iji.,lalUl 186. ll) 1"'217Fan... J- 1(4fa!oC.;nuiuo, Jl( ench:mtmem 23. 31. 39I'lTIlandez, J- 153/rh'" .86f~ti~h 59-6zligun:-grotmd 79-So, '59, 1117-8Finuk~',J. :l.fO-1Fudm-,J- A. 12;Forge, A. J1j" ... 159fl1lml:

an-rono 137--10, l18, :iJur al," p~tSOn

Fr«dbcrg-, D, 1>2 3,97, t05, IIJ n., '.fl. '50,153

Fr:ucr.J- 9. QQ-IOI. tOJ

g:l11cr)'~ 3-lJCI1dcr bias 73genealogy:

obje<:tificd IJS-40~nd reprodu<:rion of iduls 14;

Gestalt psychoklg}" 43Gibson.}.}. IfryGkUes...\. 2#Goclhe.}. W, If>eJGod 58. "-1-15. I!';, 19:1.Gohling.}. 250Gomhrich, E, 1-12, 16~, I(,R, [9iGon,orich. R, I...., ItS-50GoocIm:m, N. 25Go)'2 )5, J9

1IIIlnlll' 127IJilltrnkrrll'2 199-:00, ZIO, 21;Hal1son, F...\. H, IS<} 6:, zI6Harrison, S. Iz8Heidegger. ~1. 2HHcn~·.T. 110. IIJhicnrdtylhicrar<:hiz:aIKln;

of a'J""t/p;:lIiell! relations 23, 51. 53-4as a p.-oc-edurl' of moIiftl"3.nsrunoallon 170,

1;8, ,8" zoll, ~';-18within del~,r'l1i\e art ,6-7

lIindui,m "5, llti-ZI, Il ..-6,l.l4 SIlofstadter. D, and Dennen, O. '30Holland.]. I..Holl;" :\1. '0Iloh. E.. G. -17homunculus:

'effect" I.U-7tl' 'Iifl~,ubst"nce' of i<lol I~7

Jnda261l

OOmunculus (~.../.):lbe '.'lind·s ~I~' '30-1,222n:presenl~lionof IJ9

Jr(Jpr ldllY '74-7, H}9

Hugh-Jones, S. ~ll(1 (::'rslcns, J, 252

1311n,,1 H-5i~~>nie rcpr~"'nlatioll 25,131. 1]2imnodasm 6J.')6idenlity 219i<lols:

anioonicJiconic 97 9,110,131 2,135aninucy of 118-n, 133--'1, '42, 146-1. '49i\r~a S:amaj anitude 10.....nh '2.l~s:arl ubjecls 7, IIIconsecr:1lion of IH-SJf~>hi,>ned from trees 1.14-5ima~es as 62-"" 96as the imcriur 'mind' ufumples 136-7,1.12a, li"ing pel>l>nS 67, I~, ~. 12"'-5, I~8. '29:and Ihe \lind/Rod~ ClmtnlSl q1 S.15'--2p:rwns 2S '50-Jas proullypes ..ith primar} a~rlC)' 40:as socialagenlS 1!3 5Sioncs as 911, uS' 133'unwnpping' of 113, 145-8worshipuf 91, Iq, 1'5, 1;23-+. l.l.l-O, '+9

image::an1hropomorphic 10'),115:animacrof 118-21as a 'binding' 102, 11+de:llh of 2~-:;

u pen;on 62-"" q6;IS ~ milTOl" 120po"'cr of 62--:;.Y7, 150ali a 'sL:in' '04 6,225-7

Index/index (rcfcrcd to as 'a" objCC1', 'anworl.:'after p, 1.';.1):

in ..\gent positiun wilhin art ne~us 37,31\ 2ani,!'s Ct'1I"" as 235-6:a, <:enlrnllo the:arl nous 3b, 51,51 8. 66'choice' in ~loSipling primary ag.:ncylp"lienq

10)7;IS fC1ish 59 65mcdi:aling~' '3-15,J2.36,39 ,.0,51,

53-+.58,61-5.6; 8,83. I I'"m,-d,aling space-lime n8 32aoo.! mulTiple le"el~ of agenq wilhin 53, O[ 2as partible component of perwllS 1+. [04-6as Patic'l1l 33 5, 67"no.! process of'dc m"eri"li~~tion' ~~<)

in rda[iun 10 the ProlOllpc 35.36 7,38-<),62-5,102, '0'"

in r~blion [0 the R«ipicnt ~+. Jl ~,.j.O

as simulTaneously :\gent and Patiem jO-j

a~ tcchniallcnn for ·'""ark' 12 q:b "chiclcofprrsonhood 8,,230 r

JlUkr

Indian art 2.+, 30indi,·iduali..", J~, 139 ,.0inlemionalily:

~nd :agenc)' 6, 16-'7, ['J :20, JO, ,~'"

and mab-icand C1uOhII'I)' 100-'

philnoophical discus,in" of 122 33within the sclf-m~de inde~ -l4-5

il1lerna\istle~tcrll~listlhe",ies Qf ~'lind

112-.1.>inler-subj.:c.1i,'ity 120 ,'1"1 [82-3

Jain 118Jag:aDrulth, cuh of '-H-8J~p;lncse IandSC'dl'" arr 30lUlIg, C. 87-8

K", 2[3-14Kaeppl~r. ,\. I.. 73i"i-~ '16-8, [~i~" '1-1-.176 SoAnltI '1~-6, 212Kcplcr,J. [+hllf.. 8+-90,95Kurn. F. [6+Krnmru..:h, S. 111Kris, E, .nd Kun,., O. 39Kuchler, S, 10.\,223 SKIII~ 2~,6<)-71, 1'5.n8-.1~

1:lIfIJ"ri hi, 150-2

J.angsdorff, 0.11. mn [79,180, '141.:lM'aU1 33J.noou.;,., H. 213J.ayvd, J. 8S-8,9-l-Sl..conardo da Vinci 52-J, 56-8, 233I.hi-Str:luss, C. 9, 88, 90, ';6, ,6~. '75 n.,

19]~{', .152

rif"-sLJbslanec '~3 5linguistics and an 6 7, 1+, '~5, [0'+-5, 110Looo;, A_ 82Louis XJ\' JJ. J7I.ucrctius [05 6, roll, I 'I' 22J, HI, ~31

"""'"oonbg;ious/~ympalhelic 9,99-101~rnl crealion of anwurl.: 7' 2

,\lalangan ros, 22" 8,\Ialino\\ski. n. 9,101,229.~\alakula 90-.+.\I~ori [09.158, 160---<!, 19.+ 6,2;1 8marlet in arl 8 9!'brr. U. 16+\\aTtJuesan arrlrnal~rial cultore '0; no..tlitl IttHllti [80 4. ,88. 203, 213. ~11. 2,8material cullure ~28

.'Ia,isse.lI. 95

.'buss, M. '1,10,113, ,06, 19+lIlVC p,mcm~ 86-95MciallCSia 1+. +1, 58, 6y 72, 1"'-5, 9"', 1211,

la.' .11memory 25

and il1lel11iuIlJljuninlmliol1.11 rc"':Tpi111hlion~.13-11

as illtemar;~cd ,'isu;u im,ger~ :12.1-8,2.16Merlau I'onty, \1. 95melon,my 61--2metropo!ir..n innoentt 5Met:.<inger, J. .1+4

\Iichdangdo 18,46,59Michone, A. 78Miller,n. IIImim~sis 31, !/')-IO[, '.12, [50mimi:

a,uss ", mher n,inds lhrough arlwurl.: 15. rj,120, [.12-.1

interioril}' anu homunculi IJO-I,I.19-+O,1+1-8

:and pcrcqlIitm 33philosophical discussion of ~o, 126--)3.

23' 2,2.16 "'2resulUliOIT uf inlernali"/ettcmali~1lh\..,ri~

1.10 ',2.11 2,236-7tTap' 7'-:1,!la, 8.. , ~;;3

mil"Jcks lT8. [~r ..mirror:

'false' 3'ilNgc ~s 120'principle' in Marquc:s::an pl1stic art 11)0 rU!ie of in Bu<.klhisr conscc.T.Ilion rilts q9

.\Iodcrnism 82,If"..... LiJ4 i, 52 3. 56-8i\Iool.:crj,.." i\. and Khanna, i\I. Q5-0nwrpholugy:

and evolutionism [6<)visual "nd slrUdu!"",1 166-7,no.! transformaliun of till" nmlifs 17'-90su "lso Sl) k

,\Inrphy. H. 1,3,5-8, ,65 II.motifS:

ClLL'>e-anu-ef'fcct ",Ialions ...ilhin and bel"ccn++,10

t1"lImul.if;n~brqu\"S:Inarl 11'-90u~ also ,\brqucsan motifs undcr Iheir

indigel1culLS terms,\hnm,N·74. 11 5,22S-.,O

n~il feti,h ;[, 02l'\aR-;ssus m~lh 50'ational G,lIer~ 0~-3

'1eich. R. ,60. 2;1-8"clU1 6-;. '50 3'c..- H..brid..s 90 '"mho 187

269

obj<x:tifiCllion:and aJCt1cy 18 2',38of ...motions in 'ndc~ 31of "OCial ",lalions in index 62

(i'IIVre 153,2.12 51Oricl11:rl art 30origin;Llion of indn ",1,66-i2Orocuh 100}-15

111"'1. 186

""""""" ~12-13/NIp"" 183,186Par'}, J. "5Patient:

~rliSl ~s 28 31as CmJlllCrparl "f !\genl nas deri,ali,·... AgmT 2.\. '29index as .H 5proW[)PC a, 32,38RCt;picllI as 31-~. 52~l1d ~df-reciproc>l P~lrentlAgclll +5 ;0,

"lpallerns;

anal)sis of 73-95animation uf 76 80's atlaching penon; to Ihings 7-+-5,80,83-+a.' il1d,'."'s "fcullure 160-'and m;[~c, 1\6 90as ·prntolrpc-l,'S.~· arr ISas 'unfinished business' 80-1,93-+

p:llrona~:e ,-,,~, J3, l+. 41-8, 5+-5· 59Pu~ GYI/I [39 ,.0, 1+1, 2~2I'eir~. C. s. '3, Qper=plion i8 9. ,&;pe""-'"fpcrsonhooo:

OL~ agenq 222 3in an_lile rdati"ns wilh thing-; '.1, 61-." (n.

7+,81,83-4.96. 123mnccpt\l,dizeLl in 'l11hrOIK.,rO~Y 9 ro, qo-['disltibtllcd' 20 [,83,103,10,+, roo~s equi,."knl '" ~n ob;"'ClS 5, S, 9, 12,67,

70-1, 4)6, '50, [53, 19[, 19+ 5·fr.&oCl1lI'Jt;cnelloYc:d IJ1-",0,2H 3,253:os persona '9+ ;things as persons r8 19,20. H, 8J, 4)6, 99,

12.1 4pr~llCCti\" 56--'1philu""ph~, sre mind; p>)"\"h"r"8")"phm(jll"raphy:

aJiLl agenc~ ,o2;l>; ·artiSl.!csS' cR'dtion 35~"d ",(jmrol '05innuencinll" Dud",-mp '4+and portraits -W-S0

l'icosIio, P. I, '31, 138, ,6, 3l>OClr~ "'.; (,pohu [Ro. ,~. 187

270 /ndtx 271

Pollock, jackson JJPol~m."ia:

Auslrnl Isles 1:17-~o

Mar"u~"",s 1(,5-1~o

Tahiti 1l16-ISsa.u. Maori

1Mn 18.4"",nraiLs 25, J5, .19, ¥I 50,51-.\,98-9, IOl

f'OSSCS"i"" 67Poussin.:ss. ~j. 62Price, S. 1-2•.J, -I. ~. 15j

am! l'rice. R. 7JProtutypc/prototypc:

absent in dcr"...~ti'·e indn 75u"ting upon ilSelf u ,'\g.:ntll'2liml ..s 50,,,'ali Agent 35,;2ambig~",us position in ar! I1CXU' ]6 7as 'b<l.\~-m.n· 59.62and bodily idcnlifl~"lion .. ilh indo 6l-:;,

101~J, '0+ 10;duplicated ,,'ilhin .-in~1c ino.:x 64and idols ,.0,14. yj

impri<oned wilhrn indc-\: 10land mllnesis 'OO,IOl

as P.rieTll 32, JIl ~,~o, S-I ;r~prcscm.tion of 25-6amI "olt son-.:ry -IO,6.l-, 10J, 106, "-1

ps\"l:holOV :and ~Iion of li"ingJnon-li,ing thinp

121-3P"1-eholoi,.-ica1 Sollicn0' and SlY Ie '57, I59,

,6.1and lhcories of mino.! 1l6-.l:land lh~",r;es of pcrl-.:pl;on .n, H, 77 n.,

n-y,8J

Raphxl 46Rccipienl/rttipi~"Il1 Z4

as Ihe Agcm in Ihe an nUUS 33--1, 39elieiling a!jmey from lhe aTTwork .12, 59as thC P"lient in rhc.n nexus )1-2, )9.

,8as palron 39, -1,...8assclf---m:iprocal ;\gcm -17 8,10.1

rcific:ati"n of aesthctic response ..religion:

3S ~"unler-inluiliH' belief 12JinnUCl1~e upon Re<:ipicnllrcdpimrs "f

artllllrks -'-I, Il()-~'

and r.lliooalil)/ir.... lionalil) ofanilUdc5 ')6,IlS-16, UI, 123

thcs!'mbolic/ffal 135 6sa./~ idols

relic '-IS, ",8 9[{cm!>.-.no.!l, H. \15Renoir, ,.. , -17

rcp""""lllalitm:animniqiconie 25-6, '17-'1, 1JI-las. 'binding-' '02. 11.1 n., '1-1of l.. us~lil)' +-t 5ofdime liSions 19O-6,24J 51as a 'fk(ion' <}8p-aphioll conlra$led "ilh hl'l[(\liSlic 165and mlmcsi~ '19 101"f,\lind in idols 132-7non_rcprl'SCntatiOllal art 33.76, '9'limi1sof 125.150,25'and '51,lil representalion' 156. '9J 6and s} necdochc 1Mof timo: ~J---s8

'« ./~ :H:S1hetian:scmbbnce:

and iconic rcpresenlalion 2'; 6. l" J5, J9,-IS So, Q7-X

and mlmnis '00,150Rc)'nolds,J. 52 JRichardson, .\'a~·, sa Rnl;eby Vl"lltl.SRip.lId, Ilpdnllx: .17ritual:

ofamiL1ion (fumer) 19-.10and <:aste dUlles 14,'; j, 1.j8-9coIl,,:~rationof io.!ol<lilll<lg.:S 1..3-53ullwrapping of Ow 111-1.1saaHo idols

Riun., W. IL 87-8RoI.:eb)· \'cnlJS 62-5

Solcrili...;a1 \'iclim 6" 102-~

Sahlins, M, 106Sl/llIgrllnl(l. 1(( ston~<

Same, j .•P. 82-3Sau",m~, F. d(- 2SSehin, F, 2;schonlart 54-5Schiil7~ "'. T30;:cuII'IU~ 1')0-208.df-rccipmcal rc1a1ion~ octween lermS .,6,

-10 50, j6Sl.'m:mti<'S 5,6 7so:miotics, su aesthClic:sslltU cro..·ru 1'1-' 6sign:

arbilra~ 2S'natur:l.!' 13,26IH illllI al"Sthetics, represcnlalion

Sinha,j. '17socialagenr, sa agene)Sociology 7-11,12,127sora.~~

in Pol)'nC!<ia 1000-'-Iandroleofcxu\i~<· loS,II','u q.Jlled to religiou, \\orship ,q '5\"Olt .)2, -10, 6-1, '0' -I, ,06

Sperber, D. 10spirit mco.!ium a, anefaClllal pcrson/><IC;.T .gcm

7.pht repn-SClllarion IS6, '93-6Ion den Sleinen, K. 16,:;,168--220Sleiner, C. B. 8

'"~as both aniconic and iconic I) 1'f('rlilil~" slones '0,-9'~rowlh' ,tones '(1)

>IS repreknt~li"nal indexes or I:l>U< 98, J ~-I.

125-6Jalap-limn Mont::'< '-17-8

Strathern, A. 58Slnthcm, _\t. 74-- 1,.0, 1-11, ~l~Slill--Slcps 2o,:;~8

SI}lc r68-220as cultute 1<;.,-5ddinitiun of 167inw,iduaUcollecri,'C 157 9as'me'''lin!;, '59and ~-..ocdoche 161-2.. ithin 2mhnlJlO\og) and 3CStMfics 156

SOlherland, G. -t-9symholic:

inrerprelalion colllr:\~led wilh actioJl-{lriel1lcd.pprOllch 7

rcpre:scntation 2;1, :l;JlieU; IJ--I-j

S«.blI a.csIhetic:s.. Il:p~lation<ymm...~· ;-1.76,160-,.~)'ne<:doche 161,16;-8

"l'~hili loll T5Tambiab, S.J. 101,1,+7, 'f')

Tanlrie art 30tall""i,,~

:mal~~ of ,\larqU6an motifs t71 ~~o

and dealh 195defending lhrcshokl oflh~ body 8(',90JS a form or'wrapping' ~IS

in India ')0

Tau,,",ig.1\1. 31,99 tOO, 102. 1D.jrcchnol~' of enchantment Ht<:mple as the idols' bOOy' 136 7,1-1'18Tbomas, N. 8, '55, 168. 25', 251 nThompson, I). IjO, '76, 20'1/" 109-12lih 1<)6,1<)<) 202, '1[" 220

lime:.rllhropollt\,';~~1 pcrsp.:t:til e upon 10-'1,

l-tS-9art objecls emhodying Icmp"",lity :226-8,

23~ 58and IJul,:lIamp's fourth dimo.-nsion ~3 -50"-ub as a space-time field lUl-.'2

ItI" 100} '2Ir.ps. amvorks", 6<), llo, l:I2tre<:,:

a[trihutcd \\'ith lif(- 121nude into idols 29-)0. I-H-5

1r~rund,MtKub

,,_~ f'<rTl JI, ,g, 2;0TUTnC'r, V. 29-)0T~1or, Sir James 9, ~8-(j, III

1I'"wJrriorehlb~ 201) II

Vali:~', P. "5 6\an Gogh, V. 3.1:'4fU 182

\ elazqua., I). 62,12Vermeer,j. 6<),72Viclorian Ill, tl'

virrgt Oll'''"'''~ 141----2,'irruosit~, 72...,h sorcn)', sa sor",,~·

\\agncr,R.140,25).256\\ ashburn. I). K. Jnd C.ro,,·e, D. W. 79-80,

160,170Wc.".,.n, Mr art; atl;"l; individualismWilde, O. .12Win~h, p, G, 126WiUg<'nstcin, I.. 126.127,,·omen:

as artists 73,85ritU:lt. of affliclion 29-.1011110<><'11 ')0

wrapping:"frh(-!M>dy in lall(Mling 21Sl,f (/U~ motif 01lto a Ihrcc...limensillnal core

11)0,201of the gods 1t Iof lhe imara ofjagg-An<llh 146--'7

Yjro Him 104 S. 117, 22)

Z~no's parado.~ Illl