Dogon Restudied: a Field Evaluation of the Work of Marcel Griaule [And Comments and Replies] Author(S): Walter E

Dogon Restudied: A Field Evaluation of the Work of Marcel Griaule [and Comments and Replies] Author(s): Walter E. A. van Beek, R. M. A. Bedaux, Suzanne Preston Blier, Jacky Bouju, Peter Ian Crawford, Mary Douglas, Paul Lane, Claude Meillassoux Source: Current Anthropology, Vol. 32, No. 2 (Apr., 1991), pp. 139-167 Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for Anthropological Research Stable URL: http://www.jstor.org/stable/2743641 . Accessed: 26/03/2011 04:12

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=ucpress. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

The University of Chicago Press and Wenner-Gren Foundation for Anthropological Research are collaborating with JSTOR to digitize, preserve and extend access to Current Anthropology.

http://www.jstor.org CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I ? I991 byWenner-Gren Foundation for Anthropological Research. All rightsreserved OOII-3204/9I/3202-0002$3.00

though some are demonstrablybetter or worse." In a similarvein, Beidelman (i989:265) concludes in his review of Clifford's(I988) The Predicament of Culture, Dogon Restudied "Some accounts are of less value than others,either because the data are unreliableor thinor because the analyses are inept." I will argue here that all three kinds A Field Evaluationof the Workof of fault can be identifiedin the classic studies of the Dogon. Marcel Griaule It is a restudyof the Dogon ofMali thatI presenthere, setting offmy data against one particular (and quite well-known)other view, that of Marcel Griaule. Given by WalterE. A. van Beek the constructednature of culture,the field critique of the data has to be performedin termsof recognizability and productivity:Do the texts producedby Griaule depict a societythat is recognizableto the researcher,offer productiveinsights into Dogon thinking,or provideuse- This restudyof the Dogon of Mali asks whetherthe texts produced fulguidance for the reproductionof Dogon culture? Are by Marcel Griaule depict a society that is recognizable to the researcher and to the Dogon today and answers the question more Griaule's cultural constructsrecognizable in the field, or less in the negative. The picture of Dogon religion presented in either to the anthropologistor to the Dogon? Do the Dieu d'eau and Le renard pale proved impossible to replicate in Dogon, when presentedwith (elementsof) the Griaule the field,even as the shadowy remnant of a largelyforgotten past. texts, recognize them as a meaningfulpart of their The reasons forthis, it is suggested,lie in the particularfield situa- thinkingand way of life? tion of Griaule's research,including featuresof the ethnographer's approach, the political setting,the experience and predilections of My answersto these questions will be more or less in the informants,and the values of Dogon culture. the negative,depending on what part of Griaule's work is in question. I will firstpresent a short synopsis of WALTER E. A. VAN BEEK is Associate Professorof Anthropology Griaule's ethnographyand some of the problems it at theUniversity of Utrecht (K.i222, P.O. Box 80.I40, 3508 TC poses; then I shall present some of my own findings, Utrecht,The Netherlands). Born in I943, he received his Ph.D. analyse theirdifferences from Griaule's, and finallycon- fromthat universityin I978. He has conducted research among the Dogon of Mali and among the Kapsiki/Higi of northernCamer- structa model of how the Griaulean ethnographywas oon and northeasternNigeria. His publications include The Kap- generated. siki of the Mandara Hills (ProspectHeights: Waveland Press, I987), (with S. Pern and B. Alexander) Masked Dancers of West Africa: The Dogon (New York: Time-Life, i982), (with J.van Baal) Symbols Griaule's Dogon Ethnography for Communication: Religion in Anthropological Theory, 2d edi- tion (Assen: van Gorcum, I984), and the edited volume The Quest for Purity: Dynamics of Puritan Systems (Berlin: Mouton, I988). Griaule published on the Dogon fromthe early thirties The present paper was submitted in final form 5 IX 90. (the Dakar-Djibouti expedition of I93I-33 [Griaule I9331) until his death in I956, and a major work coauthored by Germaine Dieterlen appeared posthu- mouslyin I965. Forpresent purposes, his Dogon ethnography may be divided into three major periods. From It has become a commonplace that ethnographiesare I93I till I948 he published descriptiveaccounts of Do- "doublymediated," shaped by the ideas and preconcep- gon life, concentratingon material culture. His these tions of both ethnographerand informants.They are "a d'etat, Masques dogons (Griaule I938, hereafterMD), is tale of two cultures-the fieldworker'sand the other's" the high point of this period,which Clifford(I983:I38) (vanMaanen i988:I38). Restudies,however, represent a calls Griaule's "documentaryphase." My critiqueis not more complex case. They comment explicitlyupon an directedat his works of this periodor those of his many existingaccount and probablyenrich the corpusof data. collaborators;indeed, theirvision of Dogon cultureco- A restudymay give a differentgeneral impressionof a incides to a significantdegree with mine. It is the later culture-opening a new vista, allowing fora plural and presentationof Dogon culture with which it is con- morebalanced interpretation.("Societies are so complex cerned. thatmore than one model of analysisis needed to reveal The second period is dominated by a book that all these intricacies" [Beidelman I989:2I5].) But a re- sparked tremendousinterest in many circles,not only studymay also show that an earlierstudy was flawed. among anthropologists,and made Griaule world-fa- Though this latterpossibility is more or less ruled out mous. In Dieu d'eau: Entretiensavec Ogotemmeli(Gri- by the extremes of the postmodernistapproach, I am aule I948, hereafterDE) Griaule uncovered a Dogon persuadedthat some views on a culture can indeed be world view, cosmology,and philosophical system of a more productive, insightful,or plausible, in short, completeness and sophistication unparalleled in any "truer,"than others. Roth (i989:567) has arguedfor a other ethnography.The old man Ogotemmeli,who is distinctionbetween rhetorical devices and epistemolog- the pivot of this book, initiated Griaule into a Dogon ical significance:"no one perspectiveis privileged,even world dominatedby a splendidcreation myth, showing

'39 I40 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

the "Nazarene"-as Griaule liked to call himselfas a excision,and the appearanceof death, all are linkedwith white man-how in Dogon culturethe storyof creation the creationmyth. Griaule, who, judgingfrom the text served as a blueprintfor all facets of society,from the alone, exertedconsiderable influence on the contentof way to cultivate a field and build a house to weaving, the talks and on the structuringof the images into a potterymaking, drumming,and smithing.In this ele- coherentsystem, speculatively linked the Ogotemmeli gantlywritten account of 32 initiatorysessions, Griaule explanationsto the signs of the zodiac. describedan intricatephilosophical network linking the Though portrayedat the time as the ultimate insight outside world with Dogon society and thought. fromthe Dogon elders,Ogotemmeli's revelationswere A short summary: In the beginningAmma took a not to be the finalword on Dogon thought.A new and lump of clay, squeezed it in his hand, and flungit from quite differentcorpus of mythand culturalexplanations him. The clay spread out over the north (the top) and was offeredto Griaule and his close collaboratorGer- then the south (the bottom)of the world.This flatearth maine Dieterlen in subsequent years. The firstmajor was a female body,with an ants' nest' as its sexual or- indicationswere publicationson Sirius,symbolism, and gan,its clitorisa termitehill. Amma triedto have inter- the harp-lute (Griaule and Dieterlen i 9 oa, b; i9 5i) and coursewith her but was thwartedby the termitehill. He an attemptat synthesis (Griaule and Dieterlen I954). excisedit, thus creatingthe fox.2Subsequent intercourse However, it is Le renard pale (Griaule and Dieterlen producedthe originalNommo3 twins fromthe firstwa- I965, hereafterRP) that provides the crucial informa- ter.They gave theirmother, the earth,fibres for cloth- tion. Here another,even more elaborate and intricate ing,thus creatingthe eightfoldspirals of the sun, a vehi- view of Dogon cultural thought,based on a different cle formoisture, wind, and tornadoes.The fox,being the creationmyth, is offered.A cosmological storyis given only non-twin,succeeded in having (incestuous)inter- in which Amma (God) througha series of minutelyde- coursewith the earth,resulting in the flowof menstrual scribedacts and movementscreates the universe,stars, blood. He was given the power of speech and prediction. moons, and planets, and all aspects of Dogon culture. Amma then created human beings fromclay with the This account, even more than Ogotemmeli's, conjures help of the Nommo, male and female; they begat the up a vision of the Dogon as greatphilosophers living and firsteight ancestors of the Dogon, who were still bi- relivingtheir own creation,at ease in a world in which sexual. the warp of daily life is interwovenwith the woof of The Nommo proceededto create weaving,cultivated mythicalcreation into a marvellous tapestryof cosmic grains,and a complicatedstairway for descending to the proportions. earth,which was representedby the granary(guyo ya) The text itself starts with an indigenous theoryof and the basket. The ancestors,trees, wild and domestic signs (pp. 6i-87) differentiatingvarious levels of visual animals, reptilesand rodents,birds and fish descended representationof mythicalreality. This theoryis highly to the earth.Then the ancestralsmith, with anvil, bel- significantfor the book fortwo reasons: (i) The storyin lows, and hammer,reshaped the granaryinto a spindle facthinges upon visual representations,a factacknowl- whorl and stole firefrom the firstNommo. The latter's edged by the authors(p. i io). (2) Their account operates retribution,a thunderbolt,set the granarysliding down froman extreme"cryptological" view of the nature of therainbow. Its collision with the earthbroke the ances- symbols (SperberI975) claiming the identityof signifi- tor'sflexible limbs into human joints. A laterinfraction ant and signifie;for example, Amma, retreatingto his against the order of the firsteight ancestors produced heaven afterhis creativework, retains the signs of the the cereal grains,the chthonicserpent, agriculture, and creation,and as long as he can destroythe signs,he can the smithy. destroy the earth (p. 507). From thereon, in a convolutedand sometimesinter- The storyis hard to summarize,for it is highlyde- nally contradictorydiscourse in which Ogotemmeliand tailed and lacks any ongoingstory line. The creationin Griaule were equal partners,the structuringof the hu- this account proceeds,again, in stages and levels. After man body,totemic priests, the cult of Lebe (the earth), a firstcreation centeringon the acacia tree has been the making of drums,weaving, land cultivation,cloth- destroyed,Amma createsthe universefrom a primordial ing and finery,the architectureof the house and the egg,in which the firsteight vibrations or creativemove- totemic shrine and the layout of the village, the cult ments occur. The resultingspiral movementin the egg of the binu, the intricacies of sacrifice,circumcision, will come to representthe seven (plus one) fundamental seeds of cultivationand, later,the organizationof man and of society. Central to this account is the notion of involuteddoubleness-each being twin,each pair coun- i. In the English translation of DE, key (termitiere)has inevitably terbalanced by half a placenta. In heaven the first become "termite mound." This particularant, however, makes its nest in a hole in the ground, which is crucial to the sexual meta- Nommos are created(Nommo is the Dogon waterspirit phor Ogotemmeli uses. According to the Dogon, under these holes [see below])in the formof or with the double ofa catfish. water can always be found. A single Nommo of the originalfour pairs revoltsand 2. The yurugu is still called "jackal" in DE, to be correctedin RP escapes fromheaven, creatingspace and time (p. I75). to "pale fox" (Vulpes pallida). 3. In spelling I shall conformto Griaule and Dieterlen's, though a His placenta becomes the earth,which is therebyde- phonologically more correctspelling is available throughthe work filed.This waywardNommo, called Ogo, is punishedby of Calame-Griaule (I965, I968). Amma with the loss of speech. He tries to create life VAN BEEK Dogon Restudied I I4I

throughintercourse with his placenta-an incestuous Griaule's Ethnographyas a Problem act thatspawns bush spirits.He ascends to heaven again, steals grains (includingthe all-importantpo [Digitaria Griaule's publications, especially DE, by far the most exilis]),a piece of the sun that Amma created as a de- captivatingaccount, had an impact farbeyond the bor- fence (formerlyOgo's placenta), the teeth of his catfish dersof anthropology.Art historians (and art dealers),ar- double, and a piece of the placenta,which becomes the chitects,philosophers, and historians of religion used moon. Amma appointsthe ant, the termite,and the spi- these insightsin theirwork. The generalpublic, espe- der as guards forOgo, grindshis placenta, circumcises cially the Frenchintelligentsia, eagerly availed itselfof him, and reduces him to a fox (renardpale), henceforth thisrich treasure of profound African thought. Psycholo- dumb but with knowledgeof the future. gists became interestedin the stirringof Dogon souls The highlightof the creationstory, as well as the focus and undertookthorough in-depth interviews (e.g., Parin, ofDogon ritualaccording to this account,is the sacrifice Morgenthaler,and Parin-MattheyI963). The book was and subsequent resurrectionof one particularNommo, translatedinto many languages,including Russian and meant to be Ogo's twin,in orderto purifythe earthand Japanese.Among otherthings it stimulatedtourism to redeem the wrongsof Ogo. It is this heavenly sacrifice the Dogon habitat,the Bandiagaraescarpment. Added to that creates the central star systemaround Sirius (Gri- a visually captivatingculture and a stunningenviron- aule and Dieterlen I95ob) of which p6 tolo, identified ment, this engagingaccount of Dogon philosophyre- as Sirius B, is central.Through the streamof blood and sultedin a continuousstream of "Nazarene" (white)vis- the dismembermentof the corpse,this sacrificegener- itorsto Sanga. ates the blacksmith,numerous altars, a multitude of RP also became a source ofinspiration for many other stars, divination, the main ritual trees, rain clouds, disciplinesand professions.A much less accessible text, birds, the griots (bards),the walu (antelope), etc. The it has had a relativelyrestricted readership, and there totemicshrines testify to Amma's feat of reassembling have been few foreigntranslations. Nevertheless, this and revivingthe Nommo, as do the ancestralhouses, the highlyesoteric tale has had its own attractionand fasci- Dogon statues, and indeed most visual representations nation,and the later revelationshave been popularized (again accordingto Griaule). by others who have tried their hand at a general "de- Finally,an ark carryingthe firsteight ancestors, gener- scription"of Dogon culturefrom Griaulean sources (e.g., ated by the Nommo, descends fromheaven and colo- Saccone I984, Beaudouin I984, Palau-Marti I957). Art nizes the now purifiedearth, bearing the pure seeds of historianshave drawnon it to interpretDogon material food plants and trees as well as domesticatedanimals. cultureand by theirvoluminous writingsenhanced the Still, the wrongswrought by Ogo have renderedthem marketvalue of the alreadymuch covetedDogon sculp- mortal,and the complexwork of Dogon ritualsand cere- tures.The interpretationof Dogon architecturespurred monies is requiredboth to commemoratethe creative by Ogotemmeli was stimulated anew (N'diaye I972). and redeemingsacrifice and to prolongtheir short life There has been a Freudian reinterpretationof Dogon on earth. mythology(Michel-Jones I978). Artistsof various per- Througha transpositionof this (much more compli- suasions have soughtand foundinspiration in these Do- cated) myth the human body itselfbecomes an image gon chronicles,among themwriters (Schierbeek), sculp- not onlyof the creationbut also ofthe existinguniverse, tors (Moore),architects (van Eyck, Haan), and painters. containingall cultivatedseeds, mirroringin its develop- Numerous films have been made, among which the ment the developmentof the Nommo. Social organiza- manyproductions of Rouch standout. The weirdestcon- tion follows the original,mythical dividing lines of the nection is with the extraterrestrialaddicts of "cos- fourelements. The same holds, Griaule and Dieterlen monautology,"who have foundespecially in the Sirius argue,for relations between kin, the territorialorganiza- tales and the account of the ark some of their "defi- tion ofvillage, ward, and homestead,etc. As in DE, these nite proofs"of alien visits to this planet (Von Daniken happeningsare said to reverberatein every nook and has drawn regularlyupon Guerrier[I9751 or Temple crannyof Dogon culture-in architecture,agriculture, [I976]). Tourism receiveda freshstimulus. the main communal rituals, divination,sacrifice, lin- Historians of religion have often drawn the Dogon eage organization,the division of labour among black- into their comparativeendeavours (e.g., Pelton i980). smithsand griots,etc. Most of them have taken the texts produced A shortoverview such as this can never do justice to by Griaule at face value, revelingin the intricaciesof the richnessand versatilityof the mythcorpus (but see mythologyand creation. In their hands, the Dogon Palau-MartiI957, LettensI97I, PeltonI980, de Heusch have begun to exemplifythe quintessentialWest Afri- I985), but it must be evidentthat the textof RP presents can or Sudanic religion,preserved better on the Bandia- a wholly new paradigmof Dogon thinking.Clearly en- gara escarpmentthan anywhereelse-a notion which visaged as the firstof a long series (the second volume echoes Griaule and Dieterlen. was to deal with the first66 days of human existence Anthropologistshave reactedmore cautiously to this on earth),this has been the only volume to appear thus continuingseries of Griaule-Dogonrevelations, though far.Although later publications drew upon this corpus many have cited the material.A special conferenceon (e.g.,Dieterlen I95 9, I962, i989), thesecond volume has this type of world view resulted in African Worlds not yet come out. (Forde I954), which contains an article by Griaule and I42 1 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

Dieterlen (I954), the first attempt at a synopsis of considerableinfluence, the revelationsof Dogon myth the "new Dogon cosmogony"and still one of the most have remainedsui generis. easily accessible. Reviews were cautious (Goody Griaule oftenasserted that Dogon cultureexemplified I967, Jackson i984), attemptingto maintain an aca- a general"Sudanic" cultureand predictedthat variants demic balance betweenacceptance at facevalue and sus- of Dogon beliefs would be found among other groups. picion. However, the Bambara (e.g., as described by Zahan The generalethnographic problem posed by the Dogon [I9741and Dieterlen [I9571) never produced an Ogotem- publicationswas already becoming clear. As described meli to reveal the deepermeanings and correlationsbe- in the two centralpublications, DE and RP, the Dogon neath the surfaceof public culture.The same holds for bearvery little resemblance either to surroundingethnic the work of other Griaule students (e.g., Leiris I948, groupsor to any otherculture in Africa.No comparable Paulme I986, Paques i964). Nor were thesedeep in- set of myths,no such intricateweb of associations be- sightsever reproducedin any otherpart of Africa;Afri- tween mythand institutions,has ever been found.The can ethnographyknows only one Ogotemmeli,only one main body of anthropologicalmonographs, predomi- Renardpale. nantlyBritish at that time and "Atlantic" in lateryears, Dogon ethnographyis anomalous not only in respect had revealeda huge numberof local culturesapparently of othercultures but also internally.Though each of the operatingon quite differentpremises. Sociological, po- publicationsdiscussed is reasonablycoherent, the three litical, and historicalfactors were emphasized in these levels of Dogon ethnography-thepublications up till accounts, ideology and especially mythology being I947, Ogotemmeli,and the Renard pale-are quite in- granteda comparativelymodest role. Several commen- consistentwith each other.Palau-Marti (I 957) addresses tatorscriticized the absence of a "material grounding" this problembut clearlydoes not succeed in reconciling forthe descriptionof Dogon culture,the lack of socio- the threeversions. Dieterlen (i989) uses the initiation economic data, and the absence ofany link betweenide- paradigmto explain this conflict: the differentmyths ology and social structure(Tait i950). Douglas (I967, representvarious levels of interpretationof and initia- i968) analysed this glaringgap between French (read: tion into Dogon "secrets." Still, even thoughsome the- Dogon) and English anthropologyin terms of a differ- matic continuitiesare discernible,two qualitativelydif- ence in academic culture. When direct confrontations ferentsets ofmyths do operate.5Thus even ifwe restrict were arrangedbetween Frenchand English anthropolo- ourselves to Griaule's work, we have not one Dogon gists, each party showed clear lack of appreciation ethnographybut three. of the other's position.4The most thoroughcriticism was formulatedby Lettens (I97I), who evaluated Gri- aule's work under the tellingtitle Mystagogie et mysti- The Settingof the Restudy fication. Clifford(i983) has analysed the fieldwork methodologyand theoreticalassumptions from a decon- Dogon ethnographyoffers one of the clearest examples structivistperspective, producing a more sympathetic anthropologyhas to offerof a paradigmanomaly. It con- and sophisticatedcritique. foundseven the highlyambiguous models anthropology All in all, what has emergedfrom the criticalreading has developedfor West Africansocieties; the difference of colleagues is an enigma. The Dogon as depicted in between Griaule's construction(s)of Dogon cultureand Griaule's publications are an anomaly in Africaneth- the ones made in adjoining parts of Africaby his col- nography,even among the otherproducts of Frenchan- leagues is simplytoo great.A restudyin the field,then, thropology.Griaule's many collaboratorsalso concen- may yield more than just another view by "other tratedon religionas a crucial explanatoryvariable. Apart scribes." fromDieterlen, who had a special workingrelationship A few caveats and considerationsare in order: Any with Griaule and carriedthe torch,Zahan, Ligers,Le- comparisonover time withinone culturestarts from the beuf, de Ganay, Rouch, Paulme, Lifchitz,Leiris, and assumptionsof recognizabilityand continuity;cultures Paques made their careers in Africanethnography and are assumed to be recognizableby differentresearchers linguistics.Most of them participatedin the earlierex- and participantsas intersubjectivelyvalid constructs. peditions,before World War II. Their ethnographicwork Produced in the triangularinteraction among partici- oftenagreed with that of Griaule in grantingsymbolism pantsand betweenthem and the interpretingresearcher, and ideology a dominantplace in cultural explanation, cultures will be recognizable to a limited extent but most of them moved in quite differentdirections only-the stance of postmodernism-but still recogniz- later. The main point here is that Griaule's initiatory able. This is most apparentwhen institutionalaccounts trajectory,as Cliffordcalls it, has neverbeen paralleled rather than general interpretations,hermeneutics, or or even approximatedby any of his students-not even, feelingsare at issue (cf.Gartrell I979), and Griaule does as we shall see, by Dieterlen.Notwithstanding Griaule's writeabout institutions.

4. The proceedings of the conferenceon Voltaic cultures (Dieterlen 5. Only on a verygeneral level can the two accounts be considered I967) offera splendid example of this paradigmatic tower of Babel, transformationsof one another, as de Heusch (I985) has shown. A the English delegates speaking about totally differentissues from structuralanalysis such as his can, however, link any number of the French. myths fromany region. VAN BEEK Dogon Restudied I I43

Cultural constructswill change over time but will re- this kind of knowledge,linked in formand contentto tain certainfeatures either as such or in the shapingof ongoingpublic knowledge,should still be recognizable the transformationprocess. If the flow of time has in its culturalformat even if it is no longerbeing trans- changed society,some aspects will have changed more mitted. than others,while in various ways the "same culture" My fieldstudy of the Dogon6 was partof a multidisci- may be rediscoveredin new settingsand forms.A de- plinary cluster of researches on the theme of human scriptionof Dogon culture frombefore World War II ecologyin the Sahel. Carriedout by membersof various shouldring some bells in I989 (orI979, theyear I started disciplines(geography, physical anthropology, social an- my fieldresearch). The Trobriandsare still recognizable thropology,archaeology), research was concentratedon fromMalinowski's publications, as are other cultures two zones: the floodplainsof the Niger Delta, with the that have been restudied. As we shall see, the same old city of Djenne as the focus, and the Bandiagaraes- holds for the Dogon. Descriptions dating frombefore carpment.My own researchhad as its centraltheme a WorldWar II still offera valuable-if dated-introduc- cultural ecological concern, the relation between reli- tion to Dogon culture. gion and the survival strategiesof the Dogon. For this There is, however,the complicationthat some of the theme,an evaluationof the workof Griaule was a neces- crucial informationwas secret. Griaule emphatically sity. As it developed, it increasinglybecame an inte- stated that, though the knowledge revealed was not gratedrestudy of the Dogon. purelypersonal-it was deemed to be at the core of Do- The locationwas selected accordingly.Griaule and his gon public culture-it was known only to a few initi- collaboratorsworked mainly in Sanga, a conglomerate ates. It representedbackground knowledge that only of ten Dogon villages built close together.An important some should know and but a few needed to know. This marketas well as administrativecentre for the Dogon was the knowledge of the initiatedand formedpart of area, Sanga is in several ways atypical.Its total popula- the fabricof Dogon thoughtas a systemof secrets.Re- tion is quite large (over 6,500 at the moment),and the discoveringthis knowledgewill not be easy; yet it must earliest Christianand Muslim influenceradiated from be possible. it. Thus, it offersat once a concentrationof Dogon cul- The problem of secrecy in culture is a complex one tureand a peculiarvariant of it. Further,Sanga has been that has as yet received little theoreticalattention (cf. influencedby tourismand in a lesser way by the tradi- Bellman i984). Of course, culturesdo contain elements tion of Griaule research.I chose a village of i,8oo people thatare not readilydisclosed to outsiders.Details ofini- (fig.i) close enough to Sanga (9 km) to shareits cultural tiationand ritualsare oftennot fullypublic knowledge; variantto a significantdegree and farenough fromit to in addition to these "official" secrets, the ubiquitous avoid excessive tourism(and researchinfluence). Part of "skeletons in the cupboard"-the shameful secrets of the research,of course, had to be carriedout in Sanga, thepast and present-are kepthidden from the outsider. workingwith informantsfrom both within and beyond Griaule states clearlythat the deeperDogon knowledge Griaule's circle. The researchstarted with a reconnais- belongsto the firstclass ofsecrets, hidden not onlyfrom sance tripin JanuaryI978 followed by fieldworkfrom outsiders but also from the majority of the popula- AprilI979 to JulyI980. I thenmade eight return visits, tion. Only a small portion of the population-Griaule each lastingfrom one to threemonths, over the next ten (I952:32-33) estimatesi5%, Lettens(I97I:55I) from years.7Though I worked in the field with geographers, Griaule's text suggests 5% -has these deeper insights. prehistorians,photographers, film makers,and writers, The question then is how secret secretscan be and yet the data presentedhere derivefrom my own research. be part and parcel of a culture.As shared meaning is a My personal views on Africansociety, as well as my crucial aspect of any definitionof culture,a secret not expectationson cominginto the field,are appositehere. sharedis not cultural,while one shared by veryfew is Cultural ecology was my centralinterest, and my gen- by definitionmarginal. eral view of traditionalculture and religion was (and More importantstill, the question of secretsraises the is) characterizedby a moderatematerialism. I expected issue of culturalintegration. Though aspects of culture Dogon societyand religionto be shaped to a significant could well be at variance with each other,in a small- degreeby the exigencies of survival in a harsh natural scale society such as a Dogon village some integration environment,in which droughtsand locusts figured,as of culture can be expected. In Griaule's texts, Dogon well as historicalprocesses such as war, slave raiding, cultureis representedas fullycoherent and integrated. colonization,and state penetration.I expected a small- Moreover,cultural secrets may correspondto elements scale village society such as the Dogon to show some of public culture,sharing ways of reasoningand meth- integrationof materialbasis, social organization,and re- ods of classification.More oftenthan not,the revelation ligion.To a certainextent, I foundwhat I expected.His- of the secret is the revelationof a symbol-of the way in which esoteric concepts are materiallyrepresented 6. For stimulatingcomments on earlier versions of the text I thank (van Baal and van Beek i985). Thus, if the secrets re- my graduate students as well as Kees Op 't Land, Peter Staples, and vealed to Griaule are part of Dogon culture,one should several French colleagues. 7. The research was financed by two grantsfrom the Netherlands be able to retracethem to some extent.Of course,secret Foundation for the Advancement of Tropical Research (WOTRO) knowledgeis vulnerable,as its repositories(in this case and by the University of Utrecht, the Foundation Dapper, Agence old Dogon men) have shortlife expectancies.However, Aigle, and Time-Life. 1441 CURRENT ANTHROPOLOGY Volume 32, Number 2, April 1991

~~~~~~~~~~~~~~~~~~~~~~~~~.. ..~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.. ...

......

41~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... FIG i The scree villagerestudy~~~~~~~...... of Tireli settingfor the

toricalprocesses and ecological pressuresdo influence I also expected to findthat Griaule's creationmyths Dogon societyand culture.However, I foundDogon reli- were highlyconstructed but not directlyrevelant to Do- gion much less focused on environmentalissues than I gon daily life. However, realizing that this expectation had expected,much more a repositoryof historicalinci- would be sharedby the majorityof the profession,I was dents than a response to historicalprocesses. Also, the also aware thatfinding the mythicaland ideologicalcor- integrationof Dogon cultureturned out to be less than pus in the fieldwould be a greatercoup than "disprov- complete; one example is the loose integrationof the ing" Griaule, and I certainlyexpected to find at least mask complex into the main body of ritual associated some bits and pieces of the myths.On the whole, I was with sacrifice.8 able to findmuch less concretematerial even than I had expected.The ideas with which Griaule and his informants of 8. In otherrespects, too, my findingsdid not corroboratemy expec- worked surfacedonly as allusions, fragments tations. For instance, relations between villages and ethnic groups ritualexpression. occurredon a much larger scale than I had expected. My firstre- Working so close to and in Sanga was in a way search experience had been in the Mandara Mountains of northern a haunting experience. Griaule is still "present" in Cameroon, among the Kapsiki/Higi (van Beek i987). There, histor- Sanga, thoughmainly as the initiatorof the firstdam. ical and social relations hardly ever extended beyond the small mountain area. Though in many ways the two societies are compa- Research is considerednormal but superfluous:"Why rable (van Beek iggob), and though having lived in another millet- write this down? All the books have already been growingarea facilitatedmy personal rapportwith the Dogon way of writtenabout us!" Dieterlen's continued research,of life,I had to grow accustomed to the overt harmonyof intervillage course, keeps the memory of Griaule fresh. In the relationsand to the greatdistances covered in migratorytraditions. CNRS research house, situated between the village Dogon culture proved to be much more open than I had expected, eager to incorporatenew techniques and to establish relations with and a hotel, remnantsof Griaule's equipment are still centers of power far away. stored. VAN BEEK Dogon Restudied I I45

Restudyof the Dogon: Some Results

In orderto set offmy findingsagainst those of Griaule, V~~~~~ I shall concentrate,as he did, on Dogon religion.Issues of social organizationwere on the whole not problem- atic. Paulme (1940) presentsa picture of Dogon social organizationthat for the most part can still be found. She depicts a deeply rooted, lineage-based patrilineal, .... . - . virilocal society in which seniority,sexual division of labour,and a strongwork ethic hold sway. Though some changes have occurredin ceremonies and rituals,they too are still there.Bouju (I984) could referto de Ganay (I937) forher description of the enthroningof the hogon; after43 years,and even in a quite differentvillage, there was no need fora new account. The same holds formost of the marriageobservances that Paulme noted in I940, thoughsome details have changed and individual devi- ance fromthe normshas increased.Most of the descriptions of the bulu festivalin Sanga (Paulme and Lifchitz I936) are still valid, as are the standardaccounts of the honorifictitles (de Ganay I941), fox divination(Paulme I937), jokingrelationships (Paulme I939),and relations between the sexes (Paulme I940, I948). In my research,Dogon religionemerged as elusive and complicatedbut withinthe rangeof knownAfrican religions. It has limited relevance foreveryday life: forexample, much of agricultureis conducted without any referenceto it. Rituals may be groupedinto two major types:sacrifice, personal or communal,and the complex of death, mask, and sigu rituals,which I shall call the mask complex. The two complexes are partiallyinte- grated.The mythsof mask and sigu originssuggest that the mask complex is more recent; it does ignoresome distinctionsgoverning Dogon social life such as seniority(esp. the sigu) and lineage organization,but of course The timing of sacrifices is determined primarily by theremay be otherexplanations for this. the yearlycalendar, though divination may indicate ad- Sacrifice 2) is performedat all levels ofsocial (fig. orga- ditionalofferings. The sacrificialcalendar is governedby nization: individual,family, lineage, clan, ward,village, the onset of the rains and the subsequent countingof and even clusterof villages with a traditionof common moons. The main sacrificialevent is the great annual descent in the village of Arou It (e.g., [Dieterlen i9821). festival,the buro6(fig. 3), held just beforethe onset of followsa patternof invocation,immolation, and com- the rains (Pemu,Alexander, and van Beek i982z, Pauline munal consumption Beek familiar from (van I983b) and Lifchitz 1936); a smaller one is held afterthe har- other of Africa.All kinds of sacrifices also parts may vest. In these communal festivalsthe various categonies be made such as by special-interestgroups blacksmiths, within the village population-old and young,men and weavers,and goatherds.In principleeach type of sacri- women,caste and non-caste-reinforcetheir mutual de- ficehas its own called Amma afterthe altar, Dogon high pendencyand complementarity. god. Ammayewe is a lineage altar; a clan altar is called Closely linked to the use of individual altars are the wageu, an age-groupaltar muno, and a village altarlebe notions of sorceryand witchcraft.The generalvalue of (Dieterlen I941). An importantdistinction obtains be- overt harmony prevents manifestationsof this kind tween individualor groupaltars and village altars. The fromerupting public!ly, as any accusation shames both latter,entrusted to the care of specificclans, serve spe- partiesand an unfoundedone results in serious loss of cificgoals and purposes: they guard against intrusions face (van Beek I 98 3a). Nevertheless,both altarsand spe- fromthe bush, promotethe generalhealth and fertility cially made objects may be used foranti-social private of the village, heal, and solemnize oaths. Some altars rituals,and witchcraft,in its specificDogon definition have gained regional renown and are visited by people as the use ofreal or supernatural"poison," is a consider- fromother villages. On the whole, the systemof altars able counterpointin Dogon culture (van Beek i 99ob).9 is a very open one; new formsof altars and sacrifices have been introducedregularly throughout Dogon his- 9. The absence of any referenceto witchcraftin the writingsof tory.Thus in a village such as Tireli at least 40 different Griauleand Dieterlenis thereforenot due to any peculiarityof typesof altars can be found. Dogon culture. I46 | CURRENT ANTHROPOLOGY Volume32, Number2, AprilI99I

*,a,_ N s; - . _& 'SEd__ .s

.:ES .!{0E_ __ _: ____;' ...... | ]< it3Sg.tFil i's _ l

_^' '. _Zi . .,,.R

_ __ _ i_ _E LJ . _ v r _ _ m s_ I_m _s!-| __ S= l__- I _ __l l _ l . | I | -

* l Ls | l ji iE,2:w.: tM_|_...... _ 'S11X_ -_ =was I _w_ , - ] _ _ | i' * s

wI.I.U . .K - _ _ .X: - _ __ _ _ -- |I__ i . _ _11111 }3 X _ sN N :.s FIG. 3. Two shamansin frontof the ward's binu shrine during the annual ceremony just beforethe rains.

Witchesare people-men or women-with innatetend- for a constant flow of energyand knowledge into the encies to poison other people. Preventive measures village (van Beek and Banga I990). Masks may be viewed againstwitches are common and easy to obtain. Still, as elements of the bush introducedinto the village. In the importanceof overt harmony and the notion of thesigu, the men ofthe villagereturn to it fromthe bush shame precludethe voicingof suspicions.As a counter- with additional wisdom and fertility(van Beek i99i). pointwitchcraft is well hidden-one of the real secrets The conceptual world of the Dogon is complicated, of Dogon society. inhabitedby many beings. Amma, the high god, called The mask complex, with its grand rituals of first the maker of the earth,is the most important,giving burial(nyiu yana), mask festivals(dama), and the 60-year life,fertility, and rain. Invocations address him firstof festivalsigu, has been widely reported(Griaule I938, all, along with his "emissaries," the starsand the birds. Dieterlenand Rouch I97I). It centresaround questions His counterpart,the chthonicI,be, is oftenrepresented of death,fertility, and complementarity.Mask festivals as a serpentand is closely linked with the gerontocracy (figs.4, 5) functionas second burials resituatingthe de- in the village. In addition to these two gods, conceived ceased as ancestors.At the same time theyoffer a venue as individuals,there are collective spiritsof greatritual for theatrical performancesduring which the village and cognitiveimportance. The firstof these is the water youthcan compete(in dancing)with their peers-behav- spiritNommo, oftenrepresented as one but then as an iour normallyfrowned upon. example of his kind: many Nommo people the waters. The symbolicinterpretation of the masks has shifted Nommo is fearedas none other,commanding open wa- throughthe variousrevelatory stages of Dogon ethnogra- ter (not rain) and lightningwith his followingof famil- phy. My interpretationof these rituals, especially the iars-crocodiles, turtles,water serpents,and sheatfish. dama and the sigu, draws on various oppositionsappar- Less personalizedand also less visualized are a number ent in Dogon thought.The village/bushopposition is of collective spirits,never named individually:yeneu, fundamental.The bush (olu) is the source of all strength yebd,atuwunu (in the literatureAndoumboulou), binu, and wisdom, fearedfor its unpredictabilitybut needed and jinu. None of these collective spiritsare ancestors. VAN BEEK Dogon Restudied 147

. ..5 ......

* . ' ::: . .::: ...... ' }!. ' i ; ..

......

FIG. 4. A mask festival: arrival of the masks.

They representvarious aspects of the bush and ecologi- ing bowls forfood)-have special niches in Dogon reli- cal features(e.g., the scree),and the main characteristics gion,such as the makingand preservingof public peace. of the mask complex are thoughtto be derived from The same holds in a differentway forthe many collec- them: the masks fromthe yenezu,the ritual language tive joking relations that characterizeintervillage and (sigu so) fromthe jinu. interethnicrelations in the region(Paulme I940). The oldest man in the village serves as hogon, ritual Summing up, Dogon religion is moderately inter- officiant(Dieterlen I94I, Paulme I940, Bouju I984). twinedwith the ecological situationand changesin the Each clan has a similarfigure (fig. 6). The ward,contain- social and physicalenvironment, as it is with the social ing two or three clans, may have a shamanic priest (in structureof Dogon villages and its changes. In general the literature"totemic priest" [de Ganay I942]), who it has provedto be highlyadaptable-incorporating new officiatesat a binu shrine on behalf of members.The elements, addressing new problems and changes in functionof healer or herbalist (jbjongunu)depends on status. The historical element is pertinent,as changes knowledgeonly. The mask complexhas its own special- in the recentas well as in the remotepast have found ists, among whom the olubaru are the most important. theirway into ritual,but considerablecultural integra- They are the ones who know the traditionsand texts tion is evidentin both the public and the malign,more in the ritual language betterthan anybodyelse, having secret,aspects of Dogon religion. learnedthem during a three-month-longinitiation at the This, of course,is just a sketchand does not do justice time of the sigu festival.It is theirduty to recite them to the greatinternal variation within Dogon religion(a at the funeralsof old men. Women bornduring the sigu full-scale descriptionis in preparation).It is obvious, have a special functionas "sistersof the masks" (yasig- however,that althoughmuch of the old ethnographyis ine), and in contrastwith the otherwomen theyare not clearlystill valid, thereis a gap betweenmy resultsand supposedto fearthe masks. Finally,the castes ofDogon Griaule's. The differencesbetween his (later)data and society-the two groups of blacksmiths,the leather- mine entail a great number of ethnographicdetails, workers,the drummers,and the segeu (makersof serv- which may be subsumedin fivestatements summing up I481 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

4 . I - L i' _

* 4L~ 4s p

...~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...E-g0''i.'..:':'_| - .

FIcG. 5. The "grandmask," the central object of mask FIG. 6. A lineage elder conferringwith his clan eldest ritual. at the ward's central men's house. what Griaule assertedbut I could not replicatedespite not the basis ofany Dogon culturalinstitution. Numeri- systematicattempts to do so. cal symbolismis presentin a verylimited form(three i. The Dogon know no propercreation myth; neither formale, fourfor female, and eightfor the binu), mainly the versionof Ogotemmelinor that of the Renard pale confinedto medico/magicalrituals. Classification of ob- is recognizableto informants.The figureof Ogo (RP) is jects, animals, and plants does not followfixed numeri- unknown;the foxas a divininganimal has no privileged cal categories.No sign system or hierarchicalordering position in mythology.That Sirius is a double star is of differentparoles (sd) or levels of knowledge can be unknown;astronomy is ofvery little importance in reli- found. gion.Dogon societyhas no initiatorysecrets beyond the 4. The crucial concept of nyama, allegedly "vital completemastery of publiclyknown texts. force,"is irrelevantto Dogon religion.The etymologies 2. The "supernatural"world of the Dogon is more di- given in DE and RP are not retraceableand seem highly verseand much more vague, ambivalent,and capricious idiosyncratic. than representedin DE or RP. The role of ancestorsin 5. Dogon society is not pervadedby religion;in fact, Dogon religionis limited; they are not identifiedwith much of Dogon social life bears little referenceto reli- the other "supernatural"beings. The binu cannot be gious matters.Twins have limitedritual importance and consideredancestors and do not occupy a centralposi- are not centralin Dogon thoughtor ritual.Blacksmiths tion in mythologyor ritual.The water spiritNommo is do not functionas cultureheroes, and neitherdo griots/ not a centralfigure in Dogon thoughtand has none of leatherworkers. the characteristicsof a creatoror a redeemer. The firstpoint, of course, subsumes most of the oth- 3. Symbolismin Dogon religionis restrictedand frag- ers. Confrontedwith parts of the stories provided by mented,carried by ideas and objects sometimes quite Ogotemmelior givenin the Renardpale, my informants differentfrom the ones mentionedin Griaule's writings. emphaticallystate that theyhave never heard of them; Bodysymbolism is not the basis ofhouse plans or of the theyare not part of theirtem ("found"),their collective layout of fieldsor villages. Cosmological symbolismis customs.This holds forthe creationas such and also for V A N B E E K Dogon Restudied I 49

...... M, ...... Uwi Mu:,U z ...... zk: ...... -tv:x FE

W 2M 2tM

M ...... :"i!"i'."..I'iiik.! yNiE

Wmg.I iEE 24 j.t t.-m

MM:

......

...... :: ......

......

"ss -M4

......

...... IM

Mt ;ifil

& Mfflkw\ MR, 24

......

MWwo k, Am 4 W4

......

FIG. Tracks of the pale fox in divination plots. its parts,such as the "egg of the world" (RP), the Sirius atorypowers of the fox are not all that remarkable:in tale (RP), the original granary(DE), the creative roles theirview any animal of the bush knows the future.The of the binu, Nommo, and Amma, and numerous other fox is only one of these, and not even the "strongest." details.With the demise of the creationmyth the whole If anything,the key (black ant, with whom the masks structureof explanation and correlationin the two texts originated[Griaule I9381), and the ta (hyena) "know dissolves. more"; the ant, however,is too small forits tracks to The same holds for the pair constitutedby fox and be interpreted,and the hyena lacks "governability"(van Ogo, the principalactors in the cosmological RP. Not Beek and Banga I990). Moreover,historically conscious only are the myths totally unknown to Dogon infor- as the Dogon are, they insist that divinationwith the mants,but so is the figureof Ogo as a supernaturalbe- fox (fig.7) is a relativelyrecent practice, having reached ing.The termogo is veryimportant in Dogon, implying the area byway ofthe village ofIreli in the second halfof wealth,seniority, and power(Paulme I940, Calame-Gri- the igth century.Cowry-shell divination, still routinely aule I968). The oldest man in the village is the ogo (in practised,is consideredthe olderform. The generalterm theliterature hogon) of thatvillage, the ritualintermedi- for divination (armanga) primarily indicates cowry ary between the village and the supernaturalworld. shells, not the fox. Traces of sacred kingship linger round his person (de Is Sirius a double star? The ethnographicfacts are Heusch i985), but he is never conceived to be in any quite straightforward.The Dogon, of course, know Si- way related to a supernaturalbeing of that name. The rius as a star (it is afterall the brightestin the sky), old man is in factclosely relatedto anotherspirit that is calling it dana tolo, the hunter'sstar (the game and the offar more importance to the Dogon, Lebe, the chthonic dogs are representedby Orion's belt). Knowledgeof the earthspirit (who does not featureat all in RP and DE). stars is not importanteither in daily life or in ritual. The fox,as a divininganimal, is not consideredthe in- The position of the sun and the phases of the moon are carnationof some otherprimal being. A mythexplaining more pertinentfor Dogon reckoning.No Dogon outside his divinatorypowers referssimply to a long-drawn-out the circle of Griaule's informantshad ever heardof sigu wager between him and Amma, never to any formof tolo or p6 tolo, nor had any Dogon even heard of eme preexistenceof the foxas such. Forthe Dogon, the divin- ya tolo (accordingto Griaule in RP Dogon names for I50 I CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

Siriusand its starcompanions). Most important,no one, a periodof threemonths not in secretsbut in traditions even within the circle of Griaule informants,had ever everybodyknows. Their task is to recite the relevant heard or understoodthat Sirius was a double star (or, myths in the ritual language on special occasions accordingto RP, even a tripleone, with B and C orbiting (mainlythe funeralsof lineage elders).Their knowledge A). Consequently,the purportedknowledge of the mass does not go beyond public knowledge. As informants of SiriusB or the orbitingtime was absent.The schedul- stated,"If it is in the tem, everybodyknows it." How- ing of the sigu ritual is done in several ways in Yugo ever,these specialistsknow the corpusof public knowl- Doguru, none of which has to do with the stars.10 edge verbatim,by heart, more "correctly" and com- Of course, the Dogon do have myths.These were re- pletely than other people, and can recite the myths portedin the pre-i947 publications and consist of the withoutfault or hesitation.They receive theirinstruc- Dogon variantof the Mande myth(Dieterlen I94I), re- tion not fromtheir predecessors(who, being 6o years latingthe trekfrom Mande to the escarpment,the divi- older,are eitherdead or too old to give any instruction) sion of authorityamong the differentDogon groups,the but fromthose elderswho because of theirown interest primacyof Arou, the orderof seniorityin theirarrival and motivationexcel in masteringboth the ritual lan- on the escarpment,and their later dispersal. Another guage and the relevanttexts. corpusof mythexplains the separationof rainyand dry Thus, the secretsof Dogon societyare not at all of the seasons (as the outcome of a strugglebetween Amma initiatorykind. The knowledgedefined by the Dogon as and Lebe [Griaule I9381) and other specificfeatures of secretis, in fact,of the "skeleton in the cupboard"vari- the physical environmentsuch as rocks, springs,and ety. The best-guardedsecrets in Dogon society pertain specificcliff formations. The most importantmyth cor- to facts that shame them as membersof theirfamilLs pus consists of those connectedwith the masks and the or lineages, such as divisive past quarrels, or to the sigu; the versions that can be heard nowadays forma mechanisms and trappingsof witchcraftand sorcery. variant of the ones recordedin MD, with some omis- Shame (dogo) being a crucial concept (neglectedby Gri- sions and additions.They tell how the masks werefound aule), the less wholesome aspects of Dogon cultureare (by a woman) and became an integralpart of (male) Do- continuallyswept underthe carpetin dealingswith for- gon life.The storyof the firstsigu initiateis important eigners.Even among themselvesthe Dogon hardlytalk and is retold on ritual occasions in the ritual language about past quarrelsor presentsuspicions of witchcraft. (sigu so) that is partof this complex."1 It would be shamefulboth for them and forany audience By far the largest corpus of tradition,however, con- to talk about undesirable things (van Beek I983a). As sists of song texts,especially the song cycles that form the paramountvice in Dogon cultureis false accusation, part of funerals.The baja ni and sembee ni, which causing undeservedloss of face for the wrongfullyac- make forten solid hours of continuous singing,form a cused (van Beek iggob), people are verycareful with spo- splendid example. This collection of texts is attributed ken words. to a blind itinerantsinger who roamed the area in the If thereare no creationmyths, a differentview of the second halfof the igth century.The textsmay be semi- supernaturalbeings is required.In the DE the binu and mythicalor quite down-to-earth,and singingis an inte- Nommo, in the RP Amrna,Ogo, Nommo, and the fox gralpart of ritual.'2 are the central players in the creation drama. The ab- Neither the mythsnor the song texts-though they sence of some importantfigures, especially the yene, are sacred-are secret.In fact,the tem is public knowl- atuwunziu,yeba, and jinu (the various sortsof collective edge. When confrontedwith Griaule's tales, Dogon el- spirits),is remarkable.The binu in Dogon religionare ders remarked,"We have seen the sigu, and would we spiritsthat are referredto only as a collective (the num- not know it?" or,facetiously, "The people who said that, ber eightis indeed associated with them; in theirhonor- were theyby any chance presentat the creation,or did ific title they are said to come and drinkin groups of they come beforeit?" Every Dogon knows the myths eight),and they are highlyambiguous in nature. They and parts of the songs, thoughnot everyonecan tell or can possess people (the shaman is binugeju)but cannot sing them in full. In any case, not everyDogon is enti- be seen, only heard, and have no other relation with tled to tell myths,as this may be reservedeither to the the living than throughtrance. They are definitelynot old men of the lineage or to a special initiateof the sigu consideredancestors. Nor are the babinu, the totemsof (olubalu) (fordetails, see Griaule I938, Dieterlen and villages,wards, or castes, and in factthese are not even Rouch I97I). As boys,these initiatesare instructedover spiritsbut animals forbiddenas food.The most immedi- ately relevant aspect of these highly ambivalent "be- io. In Yugo Doguru a trace of Griaule's star hunting still lingers. ings" is that nobody traces descent fromthem. More- Informantssay that in olden times people knew about a star that over,they are fairlymarginal in ritual and in daily life. appeared duringthe year of the sigu, but furtherinquiry again leads to Griaule. The shamans, their representatives,have a similarly iI. These myths emphatically state that this mask-sigu complex marginalstatus in village society;they are neitherfeared is derived from Yugo Doguru. There is no feason to reject this nor sought afterand are subject to subtle ridicule.An- traditionas does Griaule (I 938), who insists on tracingit to Mand&. cestorsare indeed of some importancebut do not share I2. Griaule's tendency to restricthis informationon masks and with of the other funeralsto the mythical texts or expressions in sigu so may have theiridentity any spiritbeings. preventedhim fromseeing the relevance of this-as he called it-'- Nommo, the water spirit,does have some association 'Dogon amusement." with the binu but is a far more formidablecharacter. VAN BEEK Dogon RestudiedlI5I

Each bodyof surfacewater has its Nommo, oftena pair, various levels of pictorial representation,consonant male and female,and theyare fearedas is no othersuper- with the various levels of initiation.As I have argued, natural being. For a Sahelian people the Dogon have a these levels cannot be retraced,nor can the different remarkablefear of drowning,and tales of "the Nommo levels of initiationbe found in the differentdiagrams. who grabs people" abound. Yet it is Nommo who The termsGriaule uses forthem simplymean different figures as the central character in DE, where-as things.In his view bumo, yala, tonu, and toy represent twins-he createsmost ofDogon culture.In RP his role, stages of deeperknowledge; for my informantsbumo is thoughquite different,is at least as important.There he a continuoustrack in the sand (made by a serpent),yala is a redeemer,a primemover who throughhis deliberate the intermittenttrack of a bird,fox, or ant, and tonu a sacrificeredeems the sins and transgressionsof Ogo, the drawn picture,while toy means "truth." As with his firstcreated being. Afterhis immolation,Nommo, res- notions of differentparoles, Griaule, with his infor- urrected,serves as a culture hero. These two aspects mants, constructeda hierarchyof significationsthat of the Griaulean Nommo-the creator and the re- have no hierarchicalrelationship to one another. deemer-are nowhere to be found.What can be found The same can be shown with respectto numbers.Gri- is a fascinationwith dangerembodied in the Nommo. aule's texts are replete with numbers,as if the Dogon It may be that this fascinationgave rise to the elabora- were thoroughlycabbalistic, calculating every move. tion ofthis figure in creativeand redeemingmoulds, but Some numbers do functionas symbols, the principal in present-dayDogon culturearound Sanga, the notion ones being 3 (male) and 4 (female).Other numbersare of the Nommo as the key to creationis totallyabsent. relevant,too, in their contexts,8 and i2 among them. Symbolsare the hallmarkof both creationtexts. Gri- However,the 22 + 2 symbolismthat Griaule posits as aule evokes a pictureof a people surroundingthemselves crucial is not retraceable,nor is the 66 of RP. In short, with the symbols of their mythology,reliving their no "system" of correlatednumber symbolism can be mythicalpast throughalmost everyobject theyfabricate found.The scope of the number symbolismpresent is oruse. This is incorrect.The Dogon tendto make single- verylimited, evident in onlypersonal and "magical" rit- purpose artifacts,and this results in an elaborate array ual. Finally,cosmology does not serveas a symbolicref- of household objects; ethnoarchaeologicalresearch has erent either.In fact,the cardinal directionshave been uncoveredmore than i,ooo (Lane I986, Bedaux I987). de-cosmologized:"East" and "West" refernot to the di- The symboliccontent of these artifactsvaries, and only rectionsof sunrise and sunset but to the orientationof a minorityare used in ritual.Moreover, the specificob- the escarpment(in factnorth-east/south-west). jects describedin the Griaule textsas symboliccontain- Among the analytical concepts used by Griaule, the ers are oftenquite differentfrom those actually used in termnyama is a special case. Accordingto Griaule it ritual.A numberof symbolic objects are missing from means "vital force,"parallel to the Melanesian mana. the texts,and some items crucial to his symbolicanaly- His earliestwritings express his convictionthat a mana- sis have no symbolic value whatsoever.Items missing like concept will be relevant.It surfacesin MD and be- fromthe texts include cotton, the plants pedinge and comes more substantialin laterwritings (I940, i952) as sade1e, specificpottery, and clothing,especially men's well as in DE and RP. The problematiccharacter of the trousers.The reportedlyhuge symbolic content of the termis noted by Calame-Griaule in her Dogon-French basket (taju), the loom, one type of granary(guyo ya), dictionary(I965). Morphologicallythe worddoes not be- and the split drum/woodentrough (kbrb) is absent in longto the Dogon lexicon and may be Bambarain origin. my data, despite my proddingof informants.What is Dogon languagedoes have a similarword, ngawa, mean- evidentin both RP and DE and indeed can be foundin ing "polluted," "rotten," or "spoilt," implyingdanger the field is the general fascinationof the Dogon with and decay. Though Griaule's definitionof nyama does things-a traitGriaule commentsupon in DE but does include an aspect of danger,it is supposedto be positive, not considerin his analysis. (Of course, the same holds somethingsought after,a life-givingforce. My Dogon for Griaule, a museum anthropologisthimself.) Simi- informantsdo not recognizeit in the formor definition larly,the body or its parts may be invoked in some in- givenin the texts.For them a parallel conceptwould be stances in referenceto objects,e.g., the locks on granary panga (force),a word also used for physical and espe- doors are said to have "arms." However, this is not a cially muscularforce. However, as an analyticalconcept pervasivesource of metaphor.On the whole, body sym- the notion ofpanga is not importantfor Dogon religion. bolism plays a verylimited part in Dogon religioncom- Altars,when used fora long time,do gathersome panga, pared with many otherAfrican religions. but that is not the goal of sacrifice.Rather, it is an an- Drawings and diagrams,so prominentin the RP text, noyance,even a reason to abandon a particularaltar (van are in fact of limited importance in Dogon religion. Beek i988). When drawing, the Dogon primarilypicture masks, DE and RP are repletewith etymologiesof Dogon lex- supplementedby drawingsof animals or-sometimes emes. Associations such as the ones in DE of "four" -crops, while some decorationshave no recognizable with "sun" and "lizard" or "big" with "mother" and significationwhatsoever. Drawings of masks are often "cow"/Dr "seven"/with "cloth" and "the word"are not highlystylized. Reptiles and amphibiansthat belong to retraceableat all. Throughout,my Dogon informants the Nommo are frequentlydrawn but are not repre- were veryreluctant to reduce Dogon words to anything sentedin masks. Griaule writesthat the Dogon discern otherthan theirhabitual meanings;in fact,these Dogon I52 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

explanations tend towards precise and individual lex- cific about the field situation of Griaule's research emic identification,not fuzzy etymologies.Moreover, amongthe Dogon, and where,if not fromDogon knowl- most Griaulean etymologiesresult from a thoroughne- edge, do the elements of the tales come from? gation of the tonal system,which on the contraryis a The fieldworksituation of Griaule's researchhas been crucial lexemic and syntacticconstituent. analysed by several authors: Lettens (I97I), Clifford Dogon religionis not all-pervasive.Most of agricul- (i983), and Saccone (i984), to name just a few. Though ture,most of daily interactionbetween people, be it at theydisagree on some fundamentalissues, all are more the familylevel or elsewhere,is lived out without any or less critical of Griaule, pointingat some easily ob- referencewhatsoever to religious matters. There is a servedcharacteristics of his work.What is lackingis an tendencyto definereligious issues as an occupation for examination of the interactionof these characteristics old men: theytake care of the relationshipbetween the with Dogon norms and values. livingand the otherworld, and both othercategories of As hinted above, Dogon culture is orientedtowards adults,women and youngmen, simplydepend on them overt harmony.People may debate with great passion forthat function. Furthermore, religion is verymuch an but have to reach consensus in the end; a lastingdiffer- open option forthose who are interestedin it and can ence is intolerable.Dogon are veryslow to correcteach be relativelyneglected by those who choose to do so. other,leaving the otherto his opinions ratherthan fur- The same sort of "secularization" pertainsto the two nishing the correctinformation. Their internal social categoriesGriaule deems of crucial importance,twins structureis hierarchical,based on seniority;the old men and blacksmiths. are the ones who know and consequentlyshould not be Of course,as in any Africanreligion, twins are consid- contradicted.Respect for age and forrelatively small age eredspecial, but theyare not the essence ofhuman exis- differencessuch as those between consecutive siblings tence Griaule makes them out to be or the originalform is essential in social interaction.This respectfor senior- of creation. Twin festivals are held and twins are re- ity combined with the tendencytowards harmonyre- spected,but afterthe festivitiesno unusual attentionis sults in a "courtesybias" towardshigh-ranking people bestowedupon them. No mythof theirorigin has been and a tendencyto restrictthe flow of informationto found.They do function,however, in several folk sto- the lower echelons. Inferiorsare deniedinformation; the ries, one of which has mythical qualities (the one ex- apparentwishes of a superiorconstrain the information plaining the origin of thunder).Characteristically, the given to him. twins in that tale exhibitmonumental stupidity, highly A second pertinentaspect of Dogon cultureis its his- amusingfor the Dogon, who, indeed, considertwins to toricalpunctuation. A new item,be it a new agricultural be intellectuallyslightly inferior since the same intelli- invention or a new explanation of a hithertounex- gence has to be dividedbetween two. plained phenomenon,is easily accepted and integrated Blacksmiths,along with the other castes, do have a into the preexistingmould. New agriculturaltechniques specificsocial niche. They are chargedwith dampening are quickly adopted (van Beek iggoa), new tools appro- conflicts,helping to maintain at least overt harmony. priatedeagerly by the local blacksmiths,while new etio- They also have specificritual functionsin the case of logical tales findtheir way into acceptedlore. For exam- stillbirthsand on some otheroccasions, but theirritual ple, in explainingthe differencesbetween white people role is restricted.In no way do theyfunction as the cul- and Dogon the followingtale was told: tureheroes of RP and DE. A father,ancestor of us all, once dranktoo much beer, and in his intoxication,slept with his genitalsex- The Origin of the Dogon Myths: An Attempt posed. Of his two sons, the youngerone ridiculedhim, at Reconstruction while the older one, walking backwards,covered his father.As a blessingthe older,respectful, son became white and rich,the youngerpoor and Dogon. Though the principalelements of the Griaule texts are not to be discoveredamong the present-dayDogon, even Thus the storyof the drunkenNoah (Genesis 9:2I-27) as shadowyremnants of a largelyforgotten past, never- has foundits way into the storiesof these Dogon, who theless Griaule and Dieterlen did gathersome confir- emphaticallydenied that this was a "white" story.Tra- matory data during their field research in the Dogon ditionalistsand Christiansunanimously declared it to area, mainly in Sanga. My expose by no means implies be Dogon: it belonged to the te?m.In many other in- that the texts are to be regardedas forgeries;nor are stances the same process was discernible:foreign ele- they the result of an overlyactive imaginationon the ments were adopted and in a single generationbecame partof the author.The data as recordedin both textsdid "traditional."Thus the alms festivalof the Muslims has stem froma particularfield situation. The question is been adaptedto Dogon culture,resulting in the "purely" how these texts were produced.In this tentativerecon- Dogon ritual of sadaka, in which almsgivinghas been struction,I accept that an anthropologicalaccount is a changed into the distributionof beer, kola nuts, and story,a tale about tales, constructedby the anthropolo- blessings.The Dogon are quite aware ofthe gaps in their gist and his collaborators-the product of a bicultural etiologicaltales, eagerto fillthem with new stories,and interaction.The quest for the originof this particular see no particularreason forany fundamentaldistinction tale about tales, then,falls into two parts: What is spe- betweenthings learned from their Dogon forefathersand VAN BEEK Dogon Restudied I I53

fromnewcomers. In theirview, hierarchyand seniority supplementingtheir informationwith data from his are based not upon the contentof the tales or upon the own notebooks, dismissing them when they failed to amount of knowledge but upon the historyof settle- complete their tasks. Of course, much of this is more ment. Consequentlythe arrivalof new techniques and idiom than field reality.Though Griaule describedhis tales can in no way upset the fundamentalrelations on dealingswith the Dogon in this way, this does not pre- the escarpment. clude his having had relations of a differentkind with To this culturecame Marcel Griaule, and afterexten- some of them,notably with Ogotemmeli.In DE the an- sive surveyshe settledin Sanga. A museum anthropolo- thropologistis the student,to be taughtby the master. gist, he was orientedto the study of material culture, Yet even in the "series of meanderingtalks" that make thoughhis real fascinationwas with religion.One ofhis up the core of DE, Griaule is verymuch presentas an goals, already stated in his earlier works,was to show active agent. thatAfrican cultures, contrary to currentEuropean (and Cliffordcogently argues that for Griaule both these especiallyFrench) opinion, enshrined philosophies equal roles are part of an encompassingdefinition of the field to the best foundin classical Greece or India: "this con- situation, that of the theatre (Cliffordi983:I39). The ceptual structure. . . reveals an internal coherence,a notions of inequality of roles, of forcedinteraction be- secretwisdom, and an apprehensionof ultimate realities tween partners,and of concealment of true selves and equal to thatwhich we Europeansconceive ourselvesto purposeslie at the verycore of Griaule's fieldworkpara- have attained" (Griaule and Dieterlen I954:83). digm. Here, Clifford'sargument can be supplemented. Other preconceptionswere part of his baggage,too. The Dogon, too, have a greaterthan average sense of One had to do with the interculturalvalidity of the no- theatre.Their culture,especially in the mask complex, tion of mana, which among the Dogon he identifiedin is very much a performativeone, in which the public the notion of nyama. Mana was an importanttopic of persona dominatesthe privateone and the main source discussionin Frenchanthropology when Griaule set out of social recognitionis a splendidperformance in rheto- forthe field.Mauss, his teacher,used it as a ubiquitous ric or dance. So, fortheir part, the Dogon blend well into explanatoryconcept, and Griaule expectedto findit. He this definitionof the fieldworkencounter, playing their was convincedthat there was a coherentand deep expla- partwith creativeintelligence. nationfor sacrifice, and he did not contenthimself with In those days, inevitably,Griaule was part of a colo- the explanationsgiven by his informants.(The process nial presence,and the white man, endowed with power of probingcan be illustratedby his batteringquestion, and prestige,ranked high in Dogon eyes. In contrastto "Who comes to drinkthe blood of the sacrifice?",one many colonial anthropologistsGriaule in no way tried of his crucial leading questions in the interviewswith to diminish this ascribed status but rathercapitalized Ogotemmeli.) His interpretationof sacrificehinges on on it. Collectingmasks and otherobjects was easy: the the notion of nyama. The use of this concept enabled objectionsof their owners could be overridden.His view him to develop a pseudo-emictheory in which sacrifice of the white man's mission civilisatrice,which pro- was hardlylinked with social structureor interpersonal voked a conflictwith Leiris (Leiris I934), must have conflict,let alone with change, but directedto the ac- servedas a rationalefor both his fieldworkand his col- cumulation of a supernaturalprinciple, Mauss's mana. lecting practice. His position of power was comple- Another leading idea was his view of a culture as a mented by a strongpersonality, with firmconvictions cryptologicalsystem. Surfacephenomena had to be in- and clear preferences; Lebeuf, though hardly criti- terpreted,made to yield ever-deeperlevels of under- cal, mentions "son opiniatrete au travail" (Lebeuf standing.Symbols were signs of a hidden languageopen i987:xxiii). Dogon who workedwith him (or should one onlyto the initiate(Sperber I 97 5), and he evolveda mul- say "for"him?) still commenton his impatience: "Gri- tilayeredtheory of symbolism to accommodatethe awk- aule was always in a hurry,allowing neither himself nor wardfact that there were several"ultimate" revelations. us any time." Though the storiesthat circulate in Sanga In Griaule's view, any cultureharbours secrets that may of Griaule's hittingan informantare probablyapocry- not be revealeddirectly to an observer;cultural defences phal, theydo characterizethe way manyDogon see him have to be breached. even now: as a forcefulpersonality, in a situation of One consequence of these assumptionsis thatGriaule undisputedpower, with a clearly expressedpreference perceived fieldworkas a military operation (Clifford for specific informationand his own ways of getting i983:I32) with the explicit goal of "penetrating"cul- at it. tural defences,forcing the revelationof mysteriesand Griaule's fieldworkorganization was characterizedby the unravelingof the codes of the adversary.Like any a long series of shortfield trips, intensive use of a lim- assault, fieldworkhas to be total, involvingmany indi- ited numberof informantsand one "informateurprinci- vidual attacks on the cultural stronghold.His DE is pal," limited command of the Dogon language, and a quite clear on this: in the openingchapters he describes multidisciplinaryor at least multipersonapproach. His the fieldworksetting as a militarycamp where his vari- research was a matter of "expeditions," focused field ous subalternsengage the "enemy" each in his or her stays with teams, each lasting up to threemonths. He fashionand with appropriatetactics. Griaule's dealings explicitlydefends this practice (Griaule I 9 5 7) as a means with the Dogon of Ogol (not with Ogotemmeli) sound of catching one's breath for the informantsas well as high-handedtoday-correcting informants on mistakes, the ethnographer.Dieterlen has followedthe same prac- I54 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

tice to this day. Informantswere usually paid, and there- mants to disappointthe researcher.Something similar forethey gained not onlyprestige from close association happenedto me when I was looking into colour terms. with the white man but also a sizable income in the Using a standard colour chart, I had my informants slack season. The researchteam operatedfrom a venue name the various colours. Though I pointedout to them close to the touristhotel of Sanga, formerlya govern- that I was looking for single lexemes, "simple colour ment rest house, later a researchhouse owned by the terms."they regarded it as theirduty to name each and CNRS, and now the home ofMadame Dieterlen.Its situ- everyone ofthe 400-oddcolours on the chart.This insis- ation outside the village perimetersof Ogol (one of the tence stemmed fromtwo sources. Given their general largestvillages of the Sanga complex) does allow forin- fascinationwith objects, there is a Dogon conviction terviewinginformants but is not well suitedto participa- that they should in fact be able to name and classify tion in village life. anythingthat comes theirway, a kind of mild cultural But then participationwas not the model of Griaule's hubristhat may lure them into ridiculouslydetailed de- fieldapproach. His confrontationalstyle, evident in both scriptions.The otherreason is that naming becomes a DE and his fieldworkmanual (I957), not only probably game: it is funto tryto findingenious-and sometimes bettersuited his personalityand the colonial situation correct-names for new objects. Afterwardsthey ac- but also was an integralpart of his definitionof this knowledgedto me that it was not at all necessary,but form of cultural contact: a mutual theatrical perfor- theydid have a good time doing it. mance, with inevitablebackstage manoeuvres and hid- Such a game must have been especially interesting den agendas. Clifford'sanalysis is confirmedby the in- when it became a secretlanguage among informants ver- formationI received frommembers of his Dogon crew sus the foreignresearcher. Here the informants'theatri- of long standing.For them Griaule was-and still is-a cal definitionof fieldwork,as of Dogon culture itself, figureof slightlymore than human proportions,with an reveals itself.Judging from the reactionsof the old Gri- uncanny insightinto the hidden thoughtsand motives aule informants,such a situation,in which theyeventu- of people and-especially-the whereaboutsof caverns, ally gainedthe upperhand, had a strongappeal forthem. masks, and statues. "Like a binuge'inu (shaman) he In the days of colonial supremacy,the chance to control went throughthe mountainsstraight to the place where the informationflow balanced the scales of power. it was hidden," a Sanga informanttold me. Using performanceto ridicule the white man, by the This confrontationalstyle, on a more epistemological way, is not at all uncommon in Dogon culture.For ex- level, renderedhis approach more "etic" than "emic." ample, there is a mask representingthe white man, He confrontedhis informantswith items, be they arti- clothed in pantaloons and shirt,the head covered by a facts,plants, animals, or stars, and expected them to huge wooden mask painted a fieryred with long wavy provide adequate informationimmediately. For in- hair, a wild flowingbeard, and a hooked nose. In the stance,in his researchinto insect classification,insects dance in Griaule's day a colonial officerwas imitated were collected and presentedto his informantswith the writingsmall moneynotes forthe audience and saluting explicit expectation that they would have a different when he receivedhis "taxes." In our postcolonial days name foreach and everyspecies. One of his informants the touristis imitated; the same mask operateswith a told me, "He thoughteach keke' (cricket)had its own wooden "camera," forcingits way throughcrowds in Dogon name, and he did not stop." And that is what he order to get a good shot. Especially interestingis the got,one name foreach and everyinsect: not just a bojo Sanga variant (MD, p. 583): here the white person sits keke' (dung beetle), with two varieties,red and black on a chair,with two Dogon sittingon the floor;waving (these are indeed differentiatedby the Dogon), but a a notebook,the "Nazarene" asks the silliest questions: "horse dungbeetle," a "donkeydung beetle," and differ- the mask of the anthropologist! ent beetles for the dung of black monkeys,elephants, This is not to say that-in the case ofthe insects-the hyenas, turtles, and-closer to home-chickens and Dogon informantswillfully misled Griaule, just that goats. Creativitycannot be denied to these informants, theycould engagein harmlessgames in which informa- as theydistinguished between a "beetle forthe dung of tion was producedthat did not exist beforehand,all the bulls" (na jinu bojo keke), a "big horse dung beetle," while clearlyconforming to the whiteman's wishes. His and a beetle wallowing only in the dung of greyhorses list of insect names, I cannot help feeling,should have (so purugubojo kekel)(Griaule I96I:222-3).13 alertedGriaule to what was happening;in fact,he him- This kind of overdifferentiation-andnaming 24 dif- self in his manual assumes that informantsare habitu- ferent species of dung beetles can surely be called ally lyingand that the researcherhas to break through overdifferentiation-canonly be produced(and believed) this resistance (Griaule I957). Yet, not a shadow of if there is a strongconviction on the part of the re- doubt clouds his pages eitherabout the validityof the searcherthat Dogon culture is virtuallylimitless and informationor about his own interpretationsand cabbal- disposesof universal knowledge. Of course,it is also the istic arrangements.This fieldtechnique-presenting in- productof the inabilityto take no foran answerand, for formantswith as complete an arrayas possible of etic that matter,an unwillingnesson the part of the infor- data in orderto elicit a presumablyendlessly detailed culturalresponse-was also used to investigateDogon I 3. In a semantic analysis of this very article Calame-Griaule knowledgeabout astronomyand human anatomy,Gri- (i987:9) calls it a "forme de compos6e extremementproductive." aule deployingstar maps and anatomical models. VAN BEEKDogon Restudied| Ii55

Griaule's criticalinformation, including all of DE and (a situation facilitatedby his status as a lineage elder RP, came froma few informantswith whom he had of Ogol). Some of the most recent interpretations(de worked intensivelyand for a long time. For Ambara, Heusch i987) stem fromhim; his death in I987 was a withY6b6n6 the main source of RP, thisis clear enough: severeblow to the ongoingresearch. Ambara, the central he was firstrecruited in I93I (RP, p. 2og), workedwith figurein the RP as bothinformant and interpreter,is an- Paulme (I940:I3) and forLeiris (I948:ix), and continued otherstory. Paulme mentionshis ingenuityand descrip- to workfor Griaule. Ogotemmeli seems to be an excep- tive acumen and marvels at his capacity for creative tion,as accordingto DE it seems as ifhe invitedGriaule invention."Ambara's informationon customs was pre- to come to him without any previous contact between cise, but the interpretationhe spontaneouslyprovided the two. However, accordingto Kogem, his interpreter of these customs hardlyever found confirmationfrom at that time, Griaule had workedwith Ogotemmelifor the otherinhabitants of the village" (Paulme I940:566, threeconsecutive stays (overthree years) before the war, my translation).'4At the time of the RP information, long beforethe famous conversationstook place. Other I950-55, Ambarawas establishedas a matureDogon informantsare reportedto have consultedwith Ogotem- elder, whose French, though far from perfect,did al- meli in theirdealings with Griaule. As a consequence, low him to work with Griaule as an interpreter.Ac- Ogotemmeli knew very well what Griaule wanted to cordingto a later interpreter,as an adult Ambara was hear. Lebeuf (i987:xxv) as a consequence, compares unwilling to share access to the Nazarene with other Ogotemmeli to Hesiod. As Lettens has suggested,the Dogon, refusingalso any translation help from the intermissionof WorldWar II, which imposed a six-year -younger-Amadingue. hiatus for the researchers,might have heightenedthe Despite the collective aspects of this creation of a tensionwhen theyreturned to the same informantsafter mystagogicalDogon culture by the interactionof re- all that time, expectingto findnew riches. searcher,informants, and interpreters,Griaule himself The combinationof the Dogon orientationto provid- is very much presentin the analysis too. The closing ing informationand Griaule's researchorganization, ap- chapter of DE, where Griaule-unaided by any proach, and personalitycreated a fieldworkculture of informant-linksthe cosmological systemproduced by the "initiated,"both Dogon and "Nazarene." The prod- Ogotemmeliwith the zodiacal signs of Westernastrol- uct was a Dogon culture geared to the expectationsof ogy(Griaule I948), may serveas an example. Dieterlen's the principalresearcher, a culturegoverned by mythand informantsand interpreterstoday say thatGriaule exag- cabbala, whose "real secrets" were known only to the gerated,even thoughthey are on the whole not overly veryfew. The result is neitherDogon nor "Nazarene" aware ofthe publicationsin question. It would be highly but a curiousmixture of the two, bearingthe imprintof improbablethat Griaule, armed with strongconvictions a European view of Africanculture while at the same as he was, would refrainfrom stressing his own interpre- time testifyingto the creativityof the Africanexperi- tationin his publications,which is, afterall, the rightof ence. It is definitelynot an individualfraud. Most of the any anthropologist.But this aggravatedthe fundamental people concerned-Griaule himself,Dieterlen, and at problem that even the people most intimatelyassoci- least to a considerable extent his close informants- ated with him often do not recognize these writings believedin it more or less implicitly. as valid or as relevant descriptionsof parts of Dogon The interpreterswere crucial in this process, as Gri- culture. aule continued to work through them. Kogem and Griaule died in I956. Since then no new revelations Amadingue were in the course of the inquiries trans- have been forthcoming,though Germaine Dieterlen has formed from translators into informants.Kogem, a continuedyearly research trips to the area. The promise youngboy at the time of the Ogotemmelitalks, did not of the titlepage of RP ("Tome i, fascicule i") has never remain an interpreterbut with the help of Griaule en- been fulfilled.Dieterlen as co-authorpublished RP after teredthe armyand eventuallyattained the rank of cap- Griaule's death. Her influenceon the book must have tain. In the latterpart of his lifehe presentedhimself as been considerable.She soughtnew revelationsbut could an initiatein Dogon thought,considering Ogotemmeli's not find anyone with the same "cosmological exper- revelationssacred truth.He liked to view and present tise." Justbefore his death,Amadingue confidedto me himselfas one of the old men of the Dogon, one of the that it was impossible to findanyone who knew a cre- veryfew wholly initiated(he died in Bamako in i987). ation myth.Discussing the situationwith Dieterlen,he Thus, he neverfelt at ease with the cosmological myths had remarkedthat good informantscould not be found recordedin RP, which stemmedmainly from or through but had to be trained (elve). He was quite right.But Ambara. Short of condemningthem as a construct,he with the death of both the main sources of mythcon- expresslyconsidered them peripheralat best. Afterall, structionand Griaule, the prime mover, the corpus of theydid underminehis expertise.Amadingue, serving as mythremained largely as it was in I956, thoughnot all a younginterpreter before Ambara, became informateur the Ambara tales were included in RP. Dieterlen is still principal and graduallyalso developed into an initiate; he used his considerableintelligence and knowledgeto I4. "Les informationsd'Ambara concernant les coutumes etalent help Calame-Griaule with her dictionary(i968) and so- assez exactes; mais l'interpretationqu'il fournissaitspontanement ciolingusticresearch (i965) as a full-blowninformant, des ces coutumes ne trouvait presque jamais confirmationaupres albeit with a superiorcommand of the Frenchlanguage des autres habitants du village." I56 | CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

workingon the second volume, in which the contribu- tinued to produce totallynew tales. One clear instance tions of Amadingu6are added to the corpus. The main is the Sirius story.The I950 article is significantlydif- publications from the Griaule team since i956 have ferentfrom the laterRP version.One major changeis in been either elaborations on the existing myths (Die- the drawings:very few of the drawingsin the article, terlen i962), attemptsto integrateDE and RP (Palau- thoughcrucial forthe understandingof Sirius,recur in MartiI957, DieterlenI989), or additionalethnography RP. Furthermore,when comparable drawingsdo reap- (Dieterleni982). It is worthremarking that in the latter pear theyreceive verydifferent explanations.'5 It seems book Dieterlen has returnedto the pre-Ogotemmeli that rememberinga story (and elaboratingupon it) is mode of ethnography,describing a facet of Dogon cul- much easier forinformants in an oral culture than re- turenot fullyrecorded before. This book resemblesher calling a drawingwith captions. earlywritings (up to I95 i) much morethan her handling An important source for Ogotemmeli's bricolage of Griaule's revelations and is a solid piece of ethno- seems to have been the Dogon ewene (traditionalsto- graphicdescription. ries). Dogon know hundredsof these stories (Calame- Griaule I987a), which are eithercited in public speeches or related in their own right.Among the many Dogon The Myths as Bricolage storiesthat may have inspiredOgotemmeli are the tales of the twins thatgenerated the rainbow,the miraculous Clearlythe mythsin the two textsdo not stem froman tree,and many others.A clear case is the storyof Ogo endogenousDogon tradition.Yet, even ifthey have been hiding in a woven (Mossi!) basket (RP, pp. I78, I79), constructedin a peculiarinteraction between the parties which is a well-knownewene told by the Dogon with in the research,the question of where the tales come the rabbitinstead of Ogo as its hero. As these storiesare fromremains. Whence did the creatorsof the texts de- usually distributedfar beyond any ethnic borders,tales rive at least the elements of the myths?In answering fromother groups (Bambara, Sonray, Bozo, Mossi) have this typeof question,I have to speculate. contributed,too. A few items may be derivedfrom song Firstof all, neitherDE or RP containsa real text.This texts. fact,little appreciated in the commentaries,is of crucial With Ambara,new sources and inspirationssurfaced. importance.Myths throughout the worldare storiestold Firstof all, his knowledgeof otherMalian cultures,no- to an audience. Not so here. DE is a discourse,an inter- tably the Bambara and the Sonray,was important.He change,not a mythicalcreation story.Similarly, RP is spent considerabletime outside Dogon country,partly not a text at all-my account of it has simplymade it in Bamako, the capital of Mali, and used some models look like one. In fact,it consists of a sustained ethno- that may be traced to these cultures. Indeed, Griaule graphiccommentary on signs, symbols, and drawings and Dieterlenused the self-sameDogon informantsand interspersedwith isolated statementsfrom informants interpretersin their researches in these neighbouring and borrowingsfrom folktales. The authorsthemselves cultures (RP, p. 9). Ambara's French education-his acknowledgethe absence of a storyline (RP, p. I5) but study abroad as well as his schooling in Sanga--fur- do not seem to realize its implications:without a story nished him with yet furthercultural models. thereis no myth. One major source forthe RP, however,is not African In any case, as both DE and RP must have emerged at all. An importantpart of it, as well as crucialconcepts fromsustained bicultural interaction,several ways of used in the text,stems fromthe Bible. The conceptof a constructingthem and various sources must be consid- creationab nihilo, the creativeword, the two creations, ered. On the whole the later revelations(RP) seem to the rebellion of Ogo, the concept of atonement (by contain fewerDogon and more foreignelements than Nommo), the crucifixion(of Nommo), the eightpeople the earlierOgotemmeli ones. DE is mainly constructed saved in an ark are just the beginning.The storyof the out ofscraps of diverse origin. Some elementsare clearly redeemingsacrifice of Nommo is even closer to the Bi- traditionalDogon: the ants' nests,the centralrole of the ble: Nommo is sacrificedstanding upright, arms out- key (ant), twins as a special category(though not that stretched,tied to a tree (with horizontalbranches) with special), the association of binu with the numbereight, iron: duringhis sufferinghe thirstsand is offereda cup the numbers three and four as symbols for male and of water,which he spits out. Afterhis death (and dis- female,etc. More generally,the preoccupationwith sex- memberment)he is resurrectedby Amma and as a resur- ualityand the fascinationwith objects are genuinelyDo- rected being leads the continuingcreation of mankind gon, as is the importanceof speech and the metaphoric (RP,pp. i83-90). It is astonishingthat the biblical origin usage of "the word." Though the gap between DE and of these tales has never been noted eitherby Griaule'6 RP is very large,the method of reasoningis much the same. A continuedfascination with twins, an elabora- tion of number systems,and the free-floatingassocia- I5. For instance, fig. v of Ig5oa resembles fig. 86 of RP, p. 252, tion with any object in the Dogon spherethat character- where it signifies something totally different("separation of the ize DE are foundthroughout RP as well. It is significant twins"), while fig.xii of Ig5ob more or less reappears in RP, p. 63, fig. 2, as "the table of Amma," as well as in RP, p. I27, fig. 28, as that the style of "bricolage" is verymuch the same but "hibiscus." the net result, the "message," entirelydifferent. The i6. For the Bambara myths recorded by Dieterlen the Near Eastern culturalinteraction of Griaule and his informantscon- origin has been commented upon by de Heusch (i985:i68). VAN BEEK Dogon Restudied I I57

or by Dieterlen, who is well-versedin Protestantism. I979). Though this is possible, I thinksomething quite The influenceof Bible stories can also be tracedin the differenthappened. According to Amadingue, who earlierworks. In MD Griaule notes the names ofthe first workedwith the team at the time of the revelationsby Dogon, recordingin a footnotethat alternativeswere Ambara and Yeb6ne and consulted with Ambara as "Adama" and "Hawa" (P. 46). He gives this information much as the latterwould allow, Ambara neverspoke in without any comment,seemingly taking no notice of termsof a double-starsystem. What he did speak about the biblical (or in this case quite possibly Islamic) con- always, accordingto Amadingu6,were stars of different nection.17 Some biblical motifsmay be discernedin DE, "lgenerations"18(togu; the French translation is his), too, but they are few and theirancestry is more tenta- meaning(and pointingout) two adjacentstars in the sky, tive. In any case the Dogon have had ample exposureto which were to be consideredas fatherand son to Sirius the tales of the scripturalreligions. Islam has been an as a "grandfather."These stars,as Amadingu6pointed importantinfluence for centuries in the surroundingsof them out to me, were the two other stars of the Dog the Bandiagaraescarpment-as an enemy but still as a constellation.If this is so, then Griaule must have inter- source of knowledge.In additionto Islam, the Dogon of pretedthe informationgiven by Ambara and Y6b6n6 in Sanga have had first-handcontact with Christianmis- a differentfashion, as a systemof double and triplestars. sionariessince the early I930S, in factthe same stretch The point is that Griaule himselfwas verymuch aware of time in which the anthropologistswere present(and of the astronomicalspecificities of Sirius, having read much less intermittently).Ambara had frequentedthe astronomyduring his studies in Paris. Sure enough,his Protestantmission (Sudan EvangelicalMission, predom- referencesin the RP to the Sirius systemall date from inantlyBaptist) since his early youth (though he was theseyears. Just as he linkedOgotemmeli with the zodi- never converted) and knew these stories quite well. acal signs,so it was Griaule himself,driven by his own Some ofhis kinsmennowadays remember the vivid way convictions,who transformedpo tolo and sigu tolo into in which he could tell them. a mystery,linking the astronomicaldata he helped to Yet, despite all the odds and ends picked up fromthe produce with currentastronomical knowledge. In that "tresor" of Dogon culture,from neighbouring civiliza- lightit is interestingto see the way in which the remain- tions, and from Christianityand Islam, much of the der of the Griaule circle of informantsinterpreted these myths consisted of new, innovative creations by re- stars afterhis death. Though they do speak about sigu markable people combining great intelligence with a tolo, they disagree completely with each other as to creativeimagination. They managed to create fascinat- which staris meant; forsome it is an invisiblestar that ing and coherent mythical accounts, thus conveying should rise to announce the sigu,for another it is Venus theirown worldview, stressingthose aspects thatinter- thatthrough a differentposition appears as sigu tolo. All ested the researcher,and casting offthe shackles of so- agree, however, that they learned about the star from cial reality.Besides new storyelements theirprincipal Griaule. and quite astonishingachievement is the construction Summing up, the Dogon ethnographyproduced by of a system-a close integrationof elements hitherto Griaule afterWorld War II cannotbe taken at face value. unrelated.As Father Kerdran,a French Catholic mis- It is the product of a complex interactionbetween a sionarywho has lived most of his long life in Dogon strong-willedresearcher, a colonial situation,an intelli- country,marvelled, "C'est 6tonnantcomment ils tou- gentand creativebody of informants,and a culturewith jours rejoignentleur pieds." a courtesybias and a strongtendency to incorporatefor- Finally,despite the genius of the Dogon in buildinga eign elements.The tendencytowards the creationof in- mythical edifice with bricks of very differentorigins, creasingly"deep knowledge" shows itselfmuch more the influenceof Griaule himselfin the representationof towards the end of Griaule's life, with a decreasing Dogon mythologycannot be overlooked.His influence amount of "Dogon-ness" markingthe texts.Four ethno- in both the productionof data and theirinterpretation graphicperiods can be discerned:pre-Ogotemmeli, DE, is of foremostimportance. The most hauntingand elu- RP, and post-Griaule.The firstperiod is characterized sive of all Griaulean statementswill serve as a finalex- by valid description,the second and the third by the ample: the Sirius double-starsystem, often called the creationof a culturethat createsculture, and the fourth Siriusmystery (Temple I976). How could the Dogon im- by a returnto descriptivevalidity, partly through a re- part such detailed astronomicalknowledge to Griaule? hash of the revelations,and, we may hope, by the publi- Some astronomers(Pesch and Pesch I977, Ovenden cation of RP's second volume. I976) have commentedupon it, stressingthe fact that The final question, however,must be to what extent these revelationsclosely correspondedwith the knowl- Griaule managed to remain unaware of the problems edge available at that time in astronomy.It has been outlined here. It is hard to understandhow someone suggested that an occasional astronomertouring the who warnedso eloquentlyagainst inventive informants area furnishedthe Dogon with this knowledge (Sagan remainednaive about what was happeningbetween him and his informants.The other ethnographerswho

I7. Goody(I987:I29) does acknowledgethis aspect of Dogon culture,but he attributesit to the influenceof literatecultures, the i8. JohnHawkins of BrighamYoung University ingeniously sug- long-establishedIslamic centersof Djenne (insteadof Mopti)and geststhat the Frenchwords generation and giration (orbit)may Timbuctu. have been confused. I58 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

workedamong the Dogon, such as Leiris,de Ganay, and A clear explanation of why Tireli is so typical a Do- Paulme, must be reckoned better ethnographersthan gon village that therewas no need to studythe rangeof Griaule; theircontributions still stand. intraculturalvariation is missing, as are detailed life- It is possible to judge fromhis own writingsthat Gri- historiesof his key Dogon informantsand a discussion aule mighthave preferredto be evaluated as a creative of his own biases. This is especially strangebecause he writerrather than as a painstakingand patientscientist. himself raises the issue of anthropologicalconstructs Afterall, he did choose a formatfor the DE (the RP and informantaccuracy. It looks as if no improvements formatis not specificallyhis) quite differentfrom that have been made in anthropologicalfieldwork procedures of the usual ethnography,adopting a literaryform that over the last 50 years. Only when we have the mono- had already brought him some success through his graphon the Dogon that he has promisedwill we be in Abyssiniannovel Les flambeursd'hommes. His was a a position to judge who is rightabout what. sympatheticproject, aiming at the vindication of the Africanin the eyes of the European. His primarygoal was never just to understandDogon behaviour but to SUZANNE PRESTON BLIER prove a point about Africanthought. And so he did, GettyCenter for the Historyof Art and the Humanities, thoughin a differentway than he thought;claiming to 401 WilshireBlvd., Suite 400, Santa Monica, Calif. write ethnographyhe offeredanthropology a glimpse 90401-145 5, U.S.A. I7 X 90 into the highlyintriguing territory between fact and fiction, the realm of createdcultures, European as well as Ethnographyand collage have much in common. Both African.At the rim of the science of man, he embarked are constitutedof fragmentsor pieces of a whole that, upon a veritablejourney into the realm of intercultural when assembled anew, have decidedlyvarying affinities fiction.Though the DE and RP revelationsmay not be to the original.Furthermore, as with artists,significant regardedas documents Dogons, theyare outstandingas eventsin ethnographers'lives ofteninfluence their oeu- documents humaines, tributesto the convictionsof a vres,giving rise to major changes in style,subject, and European and to the ingenious creativityof an African point of view. As van Beek points out, the changes in people. Griaule's publicationsfrom the early,generally empirical works such as Masques dogons (I938) to the later and farmore abstractphilosophical tractssuch as Dieu d'eau (I948) and Renard pa'le (i965) are very striking. Comments Van Beek's analysis touches on a number of critical themesin Griaule's work,most importantthe presence or absence of comparable contemporarytraditions and R. M. A. BEDAUX the unique insightsoffered by Griaule's "informants." AfricaSouth of the Sahara Department,National My own recentresearch (i 987) on a linguisticallyrelated Museum of Ethnology,PB 212, 2300 AE Leiden, culture,the Batammaliba("Somba") ofTogo/Benin, and TheNetherlands. 4 XII 90 inquiriesinto the fieldmethodologies of Melville J.Her- skovits (Blier i 988a, I989) suggestthat such investiga- Following on the success of Freeman's attack on the tions often offerkey insights into fieldworkperspec- classic studies of MargaretMead, van Beek presentsthe tivesand methodologies.Several factors arising as much professionalanthropological community with yet an- fromGriaule's lifehistory as fromhis fieldmethods (the other attack by a second-generationethnologist upon focus of the van Beek essay) or literaryorientations (an the originaland classic work of a first-generationone, importanttheme of Clifford'swork [I9881) strikeme as conductedsome 5o yearsago and at a verydifferent stage germaneto the discussion. of developmentof the discipline.Hindsight is of course Youthfulambition and earlyrewards clearly have sig- 20/20, but faultfinding and blame fixinga posterioriare nificantimpacts on life histories.Critical forGriaule's not the same as substantivelyincreasing the breadth, choice of directionin his Dogon work,in my view, was scope, and depth of extantethnographies. his earlyidolization of explorers.We know of this inter- In assessingvan Beek's articlewe have to ask whether est from both his I945 book Les grands explorateurs it enriches the corpus of data or expands our insights (which proved popular enough to be reissued in I948) into the functioningof Dogon culture or whether,in and his eulogyof VictorSchoelcher (I948:32) comparing contrast,it merelyreflects unrecognized changes in Do- him to Alexander the Great, who "had the dream of gon society as a result of intensive acculturationand minglingEast and West and realizedit forhimself." Gri- perhapsbias in the choice of informantsknowledgeable aule's fascinationwith explorers,it mightbe argued,led in the subject of interest.That van Beek failedto obtain him to chartan explorer'slife forhimself. Accordingly, informationconfirming that obtained by Griaule 5o he participatedin the greatestFrench exploratoryen- yearsago does not a priorimean that Griaule's informa- deavor of the century,the Mission Dakar-Djibouti,re- tion did not obtain at that time. ceiving in the end the Legion of Honor. It is interesting His article is not scientific,because he does not pro- in this lightthat his earlybook Les flambeursd'hommes vide us the informationwe need to evaluate his state- (I934), at once travelog,fantasy, and popularaccount of ments; the reader simply has to believe them or not. a tripto Ethiopia and winnerof the Prix Gringoire,ap- VAN BEEK Dogon Restudied I I59

pears to have served as a model for Dieu d'eau. It is dogons. Dieterlen's lifeand intellectualhistory will un- similar in style,syntax (including the distinctivethird- doubtedlyneed to be examined more thoroughlywith person referenceto himselfas "the European" or "the respectto Griaule in the future. Whiteman"), and subject, although it traces a track throughmythical time ratherthan geographicalspace. Griaule's comments in the introductionto Flambeurs JACKY BOUJU are of considerableinterest in situatingboth this work Universitede ProvenceAix-Marseille I, Centre d'Aix, and Dieu d'eau (I935:xi-xii): 29 Avenue Robert Schuman, I362I Aix en Provence France. x scientificpublications are generallyintended for spe- CEDEX, 27 90 cialists who are the only ones to profitfrom them, This is certainlynot the firstcritical analysis of Gri- if I may say so; scholars rarelyconsent-moreover aule's works,but what is entirelynew, important,and theydo not always have the time or the material remarkableabout it is thatit comes froma fieldethnolo- means-to writefor a generalpublic; . .. as faras gistand is the productof i i yearsof researchamong the ethnographyis concerned,a science that is above all Dogon thus "revisited" in depth. Having myselfspent livingand as the public subsidizes directlyofficial io yearsworking among the Dogon,' I am struckby the missions,it has seemed to me possible, even neces- degreeto which van Beek's analyses coincide with those sary,to descend the proud slopes of eruditionand to that I have graduallyarrived at. Given the shortageof presentto a greaternumber of readers a readablework. space, I will orientmy comment around several issues A second importantfactor was the Second WorldWar. raised by his text that seem to me quite essential. While Lebeuf (I987:xxi) discusses Griaule's participa- The diversityof Griaule's oeuvre calls fora division tion in WorldWar I in considerabledetail, he never ex- ofit into threeperiods. The firstof these,represented by ploreshis political concernsor activitiesduring the pe- the works priorto I948, is characterizedby descriptive riod of the Vichy government.Whatever these were, ethnography(in particularMasques dogons)whose data Griaule appears to have been sufficientlyaffected by remain valid today. The second correspondsto Dieu them that his subsequent writingsare strikinglydiffer- d'eau, and here van Beek analyzes verywell the condi- ent fromthe prewarones. The postwarmaterial displays tions of developmentof a body of knowledgeof which, a near obsession with rationality,order, intellectualiza- alone among Dogon, Ogotemmeli seems to have been tion, and pacifistic values. Whether this change was the repository-since neithervan Beek norI nor anyone groundedin feelingsabout his role vis-a-visthe Vichy else has ever foundanything like it. The thirdperiod is governmentor anger and frustrationat a world gone representedby the Renard pale, which remains alto- awry is not clear. Some insightinto this issue can be gether strange and entirelyunverifiable in the field,2 gainedfrom Lebeuf's commentary on Dieu d'eau's com- whateverthe Dogon regioninvestigated. parison of the Dogon to the civilization most clearly Of van Beek's I9 points of differencebetween his data identifiedwith the ideals ofdemocracy, peace, and order, and those of Griaule I can confirmthe validityof all but namely,ancient Greece (I987 :xxv,my translation):' 2-those on the role of the ancestors and the position of the forwhich contrastwith his. the Dogon "live withina cosmogony,metaphysic, and binu, my findings Amongthe importantcultural traits identified by van religionwhich puts themat the level ofancient people Beek but largelyignored by Griaule I would underline and which Christologyitself would profitby study- the obvious desire of the Dogon forcollective harmony ing." This workwas furthermorewritten with the aim and consensus that is strikingto the participantob- of "puttingbefore the eyes ofthe nonspecialistpublic, server.This fundamentalsocial value is directlylinked and without the usual scientificapparel, a work that to secrecy and to shame. The domain of secrecyis in- ordinarilyis reservedonly forscholars; it moreover deed not that of esoteric knowledgerevealed only to a pays homage to the firstblack of the WesternFedera- fewgreat initiates but thatof historical or currentevents tion to reveal to the white world a cosmologyas rich thatare consideredshameful-instances ofconflict, past as that of Hesiod." (serious transgressions,betrayals, wars, massacres) or A thirdfactor is the influenceof GermaineDieterlen, beginningwith Masques dogons but especially pro- nounced in Dieu d'eau and later collaborativeworks (of i. I haveworked with the Kolum Karamb6 Dogon of the Bandiagara which thereare many in the I950s) focused on Dogon divisionsince OctoberI979. For the past threeyears I have also and other Sudanese abstractsystems of thought.Inter- worked,in collaborationwith Bruno Martinelli, among the Gye- estingly,these works are more similarin styleand syn- sum Dogon and the Sigu6Tellem of Yatenga(Burkina Faso) and the GurumDogon of the Koro circle.This latterresearch deals tax to Dieterlen's earlierwriting (I94I) than to Masques withsettlement history and local groupformation. 2. I too have severaltimes attempted a testat Kolountanga,where theold priest of the babinu, accompanied by several old men,hunt- i. The quotationsthat appear in thiscitation seem to derivefrom ers,willingly lent himself to the tediousverification of facts that theearly French introduction to Dieu d'eau. I was not able to get evokednothing at all forhim or theothers present. What is more, a copy of this version,and the Englishtranslation published by theorigin or foundingmyths that I have been able to gatherup to OxfordUniversity Press for the InternationalAfrican Institute in now containnothing even remotely resembling the central events I965 does not carrythis introduction. and personagesof the Renardpale. i6o I CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

present(over land, women,or sorcery).Competition and ward (and subtle) way. Douglas continues: "I have said rivalryare perceivedas threateningthe ideologyof con- thatthe two worksunder review are based on the Dogon sensus and are usually criticized.Indeed, they can only officialview of themselves.But they do not seek to re- become manifeston certain special occasions-at the late informants'statements to practice.Thus they fail time ofmasked or festivedances (on the plateau and the to observeany contradictionbetween ideal and actual" Bandiagaraescarpment) or traditionalwrestling matches (p. 23, my emphasis). Is this,however, really what is at (on the S6no-Gondoplain). As van Beek has seen, what stake in van Beek's endeavour?I cannot tell. is secretis essentiallyanything capable oftarnishing the My immediatereaction to his "restudy"was to recall image ofunity and consensus that,above all, the Dogon the Mead/Freemancontroversy or the debate over the attempt to impose on themselves and others. Thus differentoutcomes ofvarious substantialand solid stud- shame,one of the most importantmechanisms of social ies of the Yanomamo. These debates took place in the control,is attachedto the name of anyonewho deviates contextof a critiqueof positivist anthropology's concep- from this norm-who "has a thoughtlessmouth" or tion of truth.The theoreticalattention to objectivity, "says thingsbigger than his mouth," as the Dogon say. authenticity,and representationexemplified by the Many customs and institutions,among themthe joking workof Cliffordhas helped us to treat"truth" as a criti- relationshipsamong clans and the conciliatoryfunction cal Westernnotion that became the focalpoint of a mod- conferredon the fourcaste groups,give ample testimony ern science demarcatedby Cartesian logic. By empha- to the existence,past or present,of serious conflicts that sising that "some views on a culturecan indeed be ... no one has forgotten,but Dogon societyis "acquainted 'truer'than others"-which could be read "my view is with shame" and attemptsto conceal fromthe visitor 'truer'than others"-van Beek appears to ignore these anythingthat runs counterto its unitaryand consensual recent debates while demonstratinghis awareness of discourse.This notion of shame, thoughnot specificto them. Apparentlythey are the result of the "postmod- Dogon culture,constitutes an essential element of its ernistapproach" that in his termsis a disqualification etiquettethat must be recognizedas seriouslycompli- in itself;unfortunately, we are not given any definition cating ethnographic study, especially of settlement of this approachor any referenceto the "extremes"of it history. that he rejects. Additional readingsof the article per- Van Beek attemptsa reconstructionof the conditions suaded me thatvan Beek has serious problemswith the underwhich Dieu d'eau and the Renardpalle were gen- notion of "truth" in general and with the critical eratedand the bricolageof which they are the product. achievementsof a so-called postmodernistapproach in Veryfairly, he shows that this is not a matterof any particular.Why is the Griaule school wrong?Because individualfraud on Griaule's part,and his well-argued the Dogon say so? Or because van Beek says the Dogon demonstrationis convincing.But he also shows that, say so? in terms of its assumptions,method, and techniques, My second commentis more particularin nature.I do Griaule's researchcan no longerbe consideredethnogra- not and cannot question the integrityand precision of phy,for they contradict point for point the methodsand van Beek's account of Dogon ethnography.However, in techniques of ethnographicresearch today. In this con- an articlethe intentionof which is both to presentDo- nection,it would be illuminatingto have the testimony gon ethnographyand to use this presentationto criticize of French ethnologistswho were present in the field and oppose work done by others,one would expect to with Griaule at the time. find referenceto and analysis of most of the ethno- Finally,I have no objection to van Beek's view that graphicwork on the Dogon. The vast amount of visual Dieu d'eau and the Renard pale are "intercultural documentationconcerning the Dogon, especially that fictions"-a matter of literaturerather than of eth- producedby membersof the Griaule school, is forsome nology. reason or otherignored in spite of van Beek's awareness of its existence.Why? Griaule himselfmade two 3smm films in I938 (Au pays dogon and Sous les masques PETER IAN CRAWFORD noirs). Not having seen either of these myself,I will Granada Centrefor Visual Anthropology,University of concentrateon the manyfilms of Jean Rouch, the major- Manchester,Manchester, U.K. I5 X 90 ityof which were made in collaborationwith Germaine Dieterlen and document the Sigui ceremonies. This Chasms have always existed,for different reasons and in work constitutesone of the most remarkableachieve- differenthistorical contexts, between French ethnologie mentsof visual anthropology.It not only contributesto and Britishsocial anthropology.The Britishreaction to Dogon ethnographybut implicitlyraises some of the the work of the so-called Griaule school in generaland questionsraised above-questions concerningtruth, au- the studyof the cosmologyof the Dogon in particularis thenticity,objectivity, and reality.Rouch himselfis per- no exception. One need only invoke Douglas's (I968) fectlyaware of some of the problemsthat van Beek ap- reviewof Griaule and Dieterlen's Le renardpale (I965) parently ignores when he says, "The fundamental and Calame-Griaule's Ethnologie et langage (I965) to problemin all social science is thatfacts are always dis- underlinewhat is at stake: "If only the high metaphysi- tortedby the presenceof the personwho asks questions. cal subtletyof the Frenchcould be allied with the low You distortthe answer simply by posing the question" sociological cunning of the Anglo-Saxon" (p. 23), she (Rouch I978:22). The major problemin van Beek's arti- writes,pinpointing the problem in a very straightfor- cle is that he is looking for the same answer to two VAN BEEK Dogon Restudied I i6i

fundamentallydifferent questions. My problemis that fictionas ethnography.For thereto be any case against it becomes a question of trust ratherthan a question Griaule the attack has to allege thathis creativeinven- of truth.In whom should I trust,van Beek or Griaule/ tions in some way bamboozled the anthropologists.The Dieterlen/Rouch?Van Beek's arguments lead me to articledescribes how Dieu d'eau dazzled the Frenchin- thinkthat Griaule and he are posing the same questions telligentsia,psychologists, art historians,writers, and and that the differencesin answers imply that one of filmmakers and virtuallycreated an industryand a tour- them is wrong. It might prove fruitfulto reread both ist tradearound the idea of Dogon culture.Such attain- Griaule and van Beek to determine-to paraphrase ment of high fashionis not somethingthat professional Strathern(I987)-to whom we should attributethe hon- anthropologistswould normallyenvy or admire.If that our of havingproduced the more persuasivefiction. was all that happened, there would be little to write about in CA. For Griaule to need to be debunkedthere has to be evidence of mistaken professionalsupport of MARY DOUGLAS his views. All that can be mustered on this score is 22 Hillway, London N 66QA, England. 8 X 90 "Anthropologistshave reacted more cautiously to this continuingseries of Griaule-Dogonrevelations, though In CA a well-knownanthropologist agrees to be put in many have cited the material.A special conferenceon the dock, with a rightof reply to criticismsthat are this type of world view resulted in African Worlds formallylevelled against his work. In this case Marcel (Forde I954), which contains an article by Griaule and Griaule has been dead and most of his studentsretired Dieterlen (I954), the firstattempt at a synopsis of the fora ratherlong time. As it stands the review proposes new 'Dogon cosmogony.'" What conference?African nothingnew: the allegation that the Dogon and the fa- Worldswas planned without any conferenceas a third mous French anthropologistcolluded to invent an ex- volume in a series with AfricanPolitical Systems and traordinarymythology that had more in common with AfricanSystems of Kinship and Marriagepublished by ancient Greek than with African civilization is very the InternationalAfrican Institute. Forde as director old and, in the natureof the case, unanswerable.Anthro- made it his business to encourage scholars fromvery pologists have a hundred ways of misrepresenting- differenttraditions of research to communicate with deliberately,unconsciously, a lot, a little.The deviation each other. fromstrict standards of field reportingby the Griaule Because it is laughableto implythat the witty,irrever- team had always been excused or explainedby Griaule's ent, and very empiricalDaryll Forde should have been passionate desire to redress the balance of West- one of those swept offhis feet by Dogon or any other ern judgementagainst Africancivilization by unearth- cosmogony,the restof van Beek's argumentinvites crit- ing a philosophical system as poeticallyrich and com- icism. It is such an odd idea to put Griaule's veryper- plex as those of the classics. An earlier decade saw sonal oeuvre to the test of fieldworkthat one has to English-speakinganthropologists fired by similar con- thinkhard about how it could be disprovedin the field. cerns about reversingthe popular judgement against As van Beek is criticalof Griaule's question-and-answer "savages"-Malinowski to demonstratethat in obeying methods,he would surelyhave to thinkup some other custom theywere not automatons,Firth that theywere way of findingout whetherthe Dogon have been as to- capable of economic calculation, Evans-Pritchardthat tally misrepresentedas he now feels obliged to assert. religiousbeliefs were not irrational. As far as we can see it seems that he was reduced to The charge against Griaule is too much creative in- readingout large portions of the Griaule-Dieterlenre- vention. But, as van Beek says, it would have been a ported mythologyto representativeDogon listeners. considerablescoop for the young fieldworkerto have From theirreplies he concludes that the Dogon do not been able to have demonstratethat his illustriousprede- know the creationmyth of Dieu d'eau or of Le renard cessorwas rightafter all. Nobody needs to be told of the pale, that the water god Nommo is not a centralfigure shortcomingsof the Griaule-Dogonworld view, but he in theirthought, that the fox is only one among many would indeed be saying somethingnew and surprising divininganimals and has no privilegedposition in their in CA if he were offeringto validate it. Who can read myths. Since he is very familiarwith Dogon, having Dieu d'eau and not recognizeit forthe individualratio- done regular fieldworkamong them for many years, cinations of a brilliant,thoughtful Dogon conversing there is no doubt that he can substantiatethese and with an eager, unsophisticatedmuseum curator?Van otherstatements. Moreover, he has alreadypublished on Beek's wish to justifyit would requireconsiderable phil- theirlanguage and religionand a major studyof religion osophical ingenuityon his part. No plodding,item-by- is in preparation. itemchecklist would do the trick,as he recognizeswhen The complaintis not that he mightbe wrongbut that he talks about ways in which the surfacemanifestation he did not think an account of his own field methods of a mythmight be anchoredin fundamentalclassifica- necessaryfor a fieldevaluation forCA. This problemof tions. a methodof disprovinga set ofpersonal Dogon interpre- The argumentand style of van Beek's review recall tations should have earned most of his attention.For the attackson MargaretMead and Carlos Castaneda. All example,he ought alsc to have reportedhere on whom threewriters gained worldwide popularityand exerted he consulted. His discussion of the place of secrecyin tremendousinfluence. Mead was charged specifically culture is wanting in several respects.Bellman's book with inadequate fieldwork,Castaneda with presenting to which he refersis more comprehensive.Apart from I62 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

secretsof initiationand secrets of shame, thereis eso- of relevanceto the historyof the discipline and the na- teric knowledge reserved to professionsand esoteric ture of anthropologicalpraxis. Despite the scepticism knowledgethat is not reservedbut that is not generally voiced byvarious anthropologists and others,the models shared.The readersof CA would surelylike to know if of Dogon society expounded in Dieu d'eau and, to a he was carefulto reach into the appropriatespecialised lesser extent,Le renardpa'le have predeterminedmuch areas ofknowledge, such as medicine,sorcery, and anti- of what has been writtensubsequently. Many of van sorcery. When Victor Turner needed instruction in Beek's substantiveclaims come to me as no surprise. Ndembu ritual he found skilled teachers with special Thus, for instance, although the objectives of my re- professionalknowledge. Van Beek would have to con- searchin the Sanga regionin the early I98os were quite vince us thathe had soughtthese in vain. That would be different,along with van Beek I foundlittle evidence for a veryinteresting discovery, not immediatelycredible. the complex but nonethelessallegedly unified symbolic Does he not know that in otherparts of Africagnostic orderingof daily life describedby Griaule. Also, I found innercircles of knowledge are protected?It is moregen- no evidence for a concept of nyama or the use of any erallyplausible that some Dogon should have esoteric numericalsystem for categorising objects and littleindi- occult knowledgethan that theyshould not. So his vis- cationthat body symbolism is relevantto the interpreta- itinglist is crucial. tion of the spatial layout of houses, compounds,or vil- One of the points of contentionis that Griaule's ver- lages. I am also inclined to agreewith van Beek that the sion of a prototypicalWest Africanculture is not found "supernatural"world of the Dogon is farmore diverse even by Griaule's studentsin otherparts of the Sudan: than Griaule suggested.Van Beek's descriptionof the it is not Sudanese, it is not even African.To accept this objectives of sacrificeis virtuallyidentical to the ac- would be to denigratesome other well-known work. counts providedby my own informants. Calame-Griaule'sreport on speech categoriescannot be While the substantivecontent of the paper is clearly so summarilydismissed. It is not true that body-house important,far more significant is van Beek's explanation microcosmshave not been foundelsewhere: Le Boeuf's ofthe lack ofcorrespondence between Dogon beliefsand Habitation fali exemplifiesjust that.It is good thatvan practices as portrayedin the later writingsof Griaule Beek approves of Denise Paulme's remarkablymodern and those encounteredin the I98os. His is the firstcri- ethnography,well ahead of its time. But thereis a wor- tique to considerfully the role of the Dogon themselves ryingsense that he does not know how importantare in the process of text creation.His argumenthere has some of his assertions.For example,if the Dogon really some interestingimplications that are touched upon do not have the concept ofnyama (impersonalforce) or only marginally,and drawing attention to them may any equivalent, how do they talk about the efficacyof help to set the agenda forfuture Dogon studies. In the magic and prayer?He should at least tell us what cir- firstplace, it is not entirelyclear whetherthe type of cumlocutions they are driven to use in default. Case bricolagethat van Beek describesarises from the specific histories of disputes settled by recourse to divination situation in which Griaule's informantsfound them- were missing in all the Griaule corpus. This British selves or is much more widespread.If the latter,then anthropologistwould be impressedif he could indicate it would be interestingto know in what other sorts of that case historiesform the basis of the interpretation contexts it appears and what its broader significance of the Dogon religionthat his book will reveal. mightbe. My own feelingis thatwhile the term"brico- Finally,the interestingquestion arises whetherall the lage" correctlyconveys a sense of the eclecticism in- work on Africancosmology has been tainted by being volved in the choice of sources, van Beek's argument influencedby the Griaule team. This would writeoff not would have been strengthenedhad he examined the just the FrenchAfricanists but the extremelyempirical powerrelationships involved in the productionof com- workof young American anthropologists and arthistori- petinginterpretations. ans such as Suzanne Blier. At this point the tables are A featureof contemporaryDogon ritualsis the recur- almost turned.Van Beek has writtenwhat he thinksis rentinvocation of a concept of "tradition"to legitimate a definitivelynegative verdict on somethingthat no one particularroutine practices and existingimbalances be- thoughtcapable ofproof or disproof,but the weaknesses tweenelders and juniors,men and women. It is precisely ofhis reportraise the question whetherperhaps his first because of their collective nature that these contexts wish mightnot be fulfilled:perhaps one could devise a providelineage elders with opportunitiesto construct bettertest that would vindicate Ogotommeli and his "public personas" forthemselves and othermembers of Europeanfriend, for, after all, morethan two reputations the lineage. The latter,because of theirmore restricted are at stake. access to ritual knowledge,are not well placed to chal- lenge the formsthese constructstake. Nevertheless,in othercontexts both youngermen and women are able PAUL LANE to negotiateaspects of theirsocial identitiesto theirad- ArchaeologyUnit, Universityof Dar es Salaam, vantage. Thus, for instance, women employ material P.O. Box 35050, Dar es Salaam, Tanzania. 9 x go cultureand the organisationof household space to con- structindividual identities that contrastwith the im- This is an importantpaper that,as well as addingto our ages ofuniformity presented in the more"public" arenas understandingof the Dogon, raises a number of issues under male control (Lane I986). Again, young men, VAN BEEK Dogon Restudied I I63

throughtheir increasing participation in tourism,have How can textscollected in such a mannerbe subjected become responsiblefor presenting Dogon societyto the to critique?Could the informants,insufficiently literate outsideworld and thus acquireda means to presenttheir in French,evaluate the published proceedingsof these own readingof Dogon culture(Lane I988). cenacles? Even in the best of cases, could theypropose It is possible, therefore,that therewas a similar edge correctionswithout seeming discourteousor incurring to the actions ofsome ofGriaule's informantsthat moti- the loss of theirinterlocutors' generous friendship? As vated theirprolonged involvement with him and influ- for the anthropologists,they are obliged to trust the enced the specificform of theirworld view. Without a texts. On returningto the field, some of us have had fullerdiscussion of the social positions of these individ- doubts:why do these mythologicaltreasures always fall uals, how thesemay have changedthrough their associa- on the same ears? Have we been unable to locate the tion with Griaule, and some indication of the broader right sources? Have we somehow by-passed all this strugglesfor authority within theircommunities at the knowledge?How are we to account forour findingsif time,one can do no more than speculate. With the pas- theydo not come fromthe "patented"informants? How sage of time and the death of the major protagonists, could we possibly subject our colleagues' informantsto this opportunityhas probably been lost. Similar pro- cross-examinationwithout offendingeither of them? cesses can, however,be observedtoday that could pro- What is more, the published materials cannot be sub- vide us with insightsinto the ways in which strategies jected to a critical reading. Far from taking the form that vary with the audience may generate alternative of verbatimtranscriptions in the native language,they readingsof Dogon culture.In this regard,van Beek's ac- combine translations of brief quotations with para- count of the active role of the Dogon in the creationof phrases,interpretations, and commentariesunaccompa- Griaule's commentarieshas not only brokenthe hegem- nied by the criticalapparatus that would permitan un- onyof these textsbut also laid the foundationsfor future derstandingof the circumstances of their collection. work. Given that we are dealing with expressions of belief, mythicnarratives, or descriptionsof rites ratherthan CLAUDE MEILLASSOUX the explanationof such and such an institution,a corpus 2, rue de Mirbel, 75005-Paris,France. 6 x go ofdata as preciseand faithfulas possible should,I think, be a requirement;but it is practicallyimpossible to dis- Van Beek's article raises a serious problem in French tinguishbetween what originatesfrom the informants anthropology:is it capable of self-criticism?That a dis- and what comes fromthe anthropologist.For my part,I cussion of the work of the Griaule school should once have givenup using such materials,even when theypro- again be broughtup by one of our foreigncolleagues is vide informationthat conformto my own theses. highly significant.I can see several reasons for this: If the materialis not presentedin its originalform, I some are due to the well-knowncentralizing academic thinkit is because one would see that,as in most socie- and universitystructures we have in France,but this is ties of this region,the mythsand beliefsdo not consti- not what I intendto discuss here.What I wish to empha- tute a coherentwhole. The coherence given to them size is that the debate is paralyzedby a subjectivization should only be considereda hypothesis,but in this case of scientificthought. we have an unformulateda priori,an unquestionedabso- The work of Griaule and his school since Dieu d'eau lute, wherebyonly those elements consideredrelevant is based on a field approach that fails to meet all the are retained,adjusted, and organized. requirementsthat he himselfformulated in his M6thode When informationis treatedas quasi-initiatory,out of de 1'ethnographie(I957). Workingon Mali, I was as- reach of the profane,criticism can only turnshort or be signedfor a numberof years to the CNRS researchgroup judgedas a personalattack. Even todayone hesitatesto coveringthe country,where I chose to study the So- raise the issue, knowing that it will damage the good ninke people. I was also in charge of administrative relationsone enjoys with colleagues. One can therefore tasks, and I had nothing but pleasant relationships onlywelcome thepublication of van Beek's work,which withmy colleagues: my criticisms,therefore, have noth- seems to me to fulfillthe critical requirementsof re- ing to do with the way I feel about them. search and which publiclyinvites a long-neededdebate At a scientificlevel, however,the researchgroup often that too many apprehensionsand susceptibilitieshave seemed to me to be more an initiatoryschool than a dismissedfrom our chapels. researchlaboratory. Access to native mythologyor religious knowledge,one learnt, could only be gained by assiduously following the teachings of specific local masters,who possessed a secret knowledge that they Reply would divulge with greatreluctance only to those who had earned their trust. Thus knowledge of the Dogon religioncame to be "confessed" by a verylimited set of W. E. A. VAN BEEK select informantsthrough whom a corpus was con- Utrecht,The Netherlands.4 XI 90 structedthat seems, on examination,not so much a source of discoveryto the researchersas one of surprise The commentsindicate that the problemof Dogon eth- to the investigated. nographyis still importantin Africanistanthropology, I64 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

representingas theydo a wide arrayof opinions. Given of a more general arena for constructionof individual a subject as delicate as an attack upon an established identities.I think he is right.For the informantstheir reputationand the emotionsgenerated in responseto it, close association with the powerfulFrenchman must several typesof reaction are to be expected. One is the have been an importantasset in the village arena and indignantreaction, aiming at preservingthe reputation changedtheir self-images. One can, indeed,perceive just of the attackedmaster. This reactionis absent here but this with the second-generationinformants working probablywill be heardlater. A second is the uneasy reac- now withDieterlen, though I thinkthe effectis tapering tion,pointing out that with one major pillar of the eth- off. nographicbuilding gone, others are in dangerof crum- Bouju in a way offersa venue for this type of study bling as well-one of the arguments Douglas uses. when he concentrateson the twin values of shame and Anotheris the reaction focused on theory,style, and secrecyin Dogon culture.His analysis of these themes presentationmore than the argumentitself. Crawford, is a welcome elaborationof my argument.The onlypos- Douglas, and Bedaux writealong this line. Finally,there sible point of disagreementbetween us mightbe in eth- is acceptance of the article's conclusions and expansion nographicdetails (binu and ancestors),though I think upon its theoreticaland methodologicalimplications; thatinternal cultural variation among the Dogon might the commentsof Blier, Lane, Meillassoux, and Bouju fall be one factoraside fromthe inevitable personal styles into this category,and I will begin with them. and interpretations. Blier bringsan interestingthought to the discussion: Meillassoux, another anthropologistquite familiar the idea of using biographicalanalysis to establishhow with the largercultural region, agrees with my analysis significantevents in ethnographers'lives may explain and uses the conclusions to illuminate an internal changes in theirwork. She points to three such influ- Frenchacademic problem,the absence of self-criticism. ences on Griaule, and this amounts to a welcome bio- His descriptionof the internal dynamics of research graphicaladdition to my argument.Griaule did indeed groupsas initiatoryschools is fascinatingand revealing. see himselfas an explorer,and several of the field ap- As he is the firstto admit,such self-criticismas he deliv- proachesmentioned fit in well with this self-definition, ers raises the question of power in the Frenchacademic forexample, the view of the otherculture as a bastion structure.Both the tendency toward person-oriented and the use of aerial photography.Still, therehave been schools and the hierarchiesinherent in the Frenchaca- explorerswith a less militaristicvision, and Griaule's demic system would be fascinatingobjects for further favorablecomparison of a colleague to Alexander the study. Great is revealing.As for Blier's second point, I agree Crawfordraises two problems, my position in the that the Second World War may have been a turning postmodernistdebate-the question of "truth"-and pointfor Griaule. Yet, despitethe evidentchange in the the relevanceof visual documentation.As forthe latter ethnographyafter the war, many of Griaule's convic- issue, the Dogon are indeed one of the most extensively tionsand drivescan be foundbefore that time, for exam- filmedgroups in Africa,and I fullyagree with Crawford ple, his insistenceon mythas an explanationfor ritual that Rouch's films,made with the collaborationof Die- and his conviction that African societies engaged in terlen,constitute one ofthe highpoints in visual anthro- "deep thought."In any case, the comment calls for a pology. For me, as for any anthropologistworking on more generalstudy of the riftsin academia in wartime Dogon culture, these films-especially the sigui France and one in which the position of scholars such documentation-constitute a crucial source of infor- as Leiriswould be importantas well. Blier's last remark, mation. However, the same problem confrontedby on the influenceof Dieterlen,was also made by review- any anthropological research is evident in visual ers of earlierversions of the article,and I may indeed anthropology-the interactionbetween two cultures. have underplayedher importancein the constructionof Judgingfrom my own experiencewith filmingand from Dogon culture.Of course,the last periodof revelations, the literature,this may be even more poignantin visual that of the Renard pale, was in factstrongly influenced anthropology,as the presence of impressive hardware by her,as most of the works of that era are coauthored. and the strictexigencies of filmingimpose a consider- One can but wonder whetherGriaule would have pub- able amount of Western culture on-in this case- lished the Renard pale in the same form,given his pro- Dogon proceedings. Moreover, the selectivity of the clivityfor writing for a more generalpublic. But even if camera,the cutting,and the commentaryhighlight the Dieterlenwas and still is the systematizer,Griaule was interactivecharacter of the documentation.This holds the motorthat generatednew revelations. forany film. In any case, the quality of Rouch's films The commentsof Lane and Bouju are of special rele- is in my view not dependenton his incidental use of vance, as theyalso have done researchin Dogon culture mythopoeticalcommentary. recentlyand in the same region.For them the lack of fit The firstpoint Crawfordraises is more complicated between the Griaule writingsand empirical data from and quite germane to the general framework.My cri- the field comes as no surprise,and for them too the tique of Griaule's work cannot be reduced to a gap be- differenceis more than one of style. Lane raises the tween the ideal and the actual, which mighthave been question whetherthe bricolagethat generatedmany of a critique of the Dogon ethnographyup to the Second the Griaule "data" resultedfrom the specificfield situa- WorldWar. It is the untraceableideal that bothersme, tion at the time and then argues that it may be part and here the notion of relativism comes in. Crawford VAN BEEK Dogon Restudied I I65

rightlyreads my argumentas-also-a partial critique One of Bedaux's argumentsruns along similar lines; of postmodernism.Indeed, I do have problemswith the forhim as forall of us the proofof the puddingis in the notionof "truth."Who does not? While recognizingthe eating.There is, however,no reason to reservethe label meritsof the postmodernistdebate, I do not thinkthat "scientific" (whateverthat means) only for extensive the notion can be done away with by a relativiststance. presentationsof data. In factmore than half of the article ThereforeI have translatedit in termsof recognizability; is data, and I also pointedout what partof the literature in principlethis implies a search not so much fortruth could be consideredvalid ethnography-which,as he is as foruntruth, and my essential claim is that Griaule's well aware,already constitutes massive documentation. Dogon tales are unrecognizablein the field.That is in- With appropriatereferences to such a body of literature deed a questionof trust, but thatCrawford feels the need a discussionis not at all un-"scientific."If this were not to choose between the two "tales" is at least an indica- so, most discussions in our journals would be out of tion thatthe differencesbetween our two "intercultural bounds. While I accept the position that only a full tale fictions"are too large to be comfortablewith. can be fullyconvincing, the details requestedby Doug- Douglas raises a numberof interestingissues that for las and Bedaux can only be presentedin the contextof a valid answer would take up much more space than is a monograph. allowed here.One ofher points is thatthe anthropologi- How many people are implicated with Griaule? Not cal communitywas not bamboozled by Griaule's revela- as many as Douglas thinks. Of course, empirical re- tions thoughother professions were. This is correctin search into symbolic and cosmological interpretations the sense that anthropologistshave reacted to them is possible and can yieldfascinating results; for example, much morecautiously. Among them (to correctan obvi- I fullyshare Douglas's appreciationof Blier's work. Still, ous misunderstanding),Daryll Forde was in no way it is a farcry from an Africancosmology to a tripleSirius swept offhis feet despite his personal friendshipwith system.The whole point is that this researchis empiri- Dieterlen; Douglas probablyis correctin assumingthat cal and, like any such work (grantedthe limitations he mightbe the last to be dazzled by these construc- of the ethnographicalaccount), is capable of "dis- tions. I just wanted to point out the greatdistance be- proof"-that is, should be recognizable.Surveying the tween the two discourses,Griaule's and the British,and implicationsof my tempereddisproof of Griaule's work one example of that distance can be found in the pro- is not an undertakingI choose to burdenthis article with ceedings of the Sonchamp conference.Douglas's argu- and would call forfurther evaluation. In any case, I see ment essentiallyreads that in criticizingGriaule I am no reason whateverto writeoff all work on Africancos- flogginga dead horse.For the anthropologicalprofession, mologyand hope to continueworking in that fieldmy- especially the British,this may be so, but I doubt it. selffor many years to come. Griaule's Dogon materialis still used as an example of A last point on the referencesto the other debates, intricatemythological reasoning and classification,even especially Mead/Freeman:I chose the CA* formatin in anthropologicaltexts, whereas, as I triedto show, its order to open a discussion; that is what the journal epistemologicalstatus cannot be equated with that of stands for.While Griaule cannot respond,his scholarly ethnographicrecords. heirs can. Fromthe startI recognizedthe personalfeel- Douglas raises a fundamentalissue when she dis- ings involved,and forthis reason I presentedthe article cusses the logic and methodologyof disconfirmation. in personin Paris and prepareda Frenchtranslation for The problemsof secrecyare appositehere. I limitedmy- Dieterlen's benefit.Whereas none of the commentators selfto those two of Bellman's fourtypes of secrecythat so far have contradictedfield data and thereforemost are relevant to this discussion. Indeed, the Dogon do have accepted the gist of the argument,defenses of the have knowledge reservedfor the professions-such as "data" are still to be expected and are said to be under the artisangroups-as well as secretsof medicine,sor- way. This is what I hoped for,and it is herethat compar- cery,and anti-sorcery,which indeedI did exploreduring ison to the way other debates have been conducted my research.Of these the firstis neithervery secret nor falls short. very esoteric,in fact being part of the public domain. The second genreis more interestingand more difficult to gain access to. In Dogon culturemedicine and anti- sorceryare partof the darkside of life,out of tune with ReferencesCited the ideal of social harmonythat Bouju elaborateson in his comment.Griaule and Dieterlen completelyignored BEAUD OIN, GERARD. I984. Les Dogons du Mali. Paris:Armand thisdomain, and these secretsare definitelynot thekind Colin. Griaule meant. as can be seen from account BEDAUX, R. M. A., AND J. D. VAN DER WAALS. I987. "Aspects Also, my ofthe life-span of Dogon pottery,"in A knapsackfull of pottery. (van Beek iggob), the notions and symbolsused in it do Editedby L. Hacquebordet al., pp. I37-53. Leiden:Leiden Uni- not fitinto the Griaulean schemes at all. Whetherone versityPress. could call this knowledgea gnosticinner circle is highly BEIDELMAN, T. I989. Reviewof: Thepredicament of culture, by questionable.A detailed account of these problems,as J.Clifford (Berkeley: University of CaliforniaPress, i988). An- ritual thropos84:2I5-i6. well as the case historiesfor the interpretationof BELLMAN, B. L. I984. Thelanguage of secrecy: Symbols and met- and symbolism,should, as Douglas suggests,and will be aphorsin Pororitual. New Brunswick:Rutgers University Press. includedin the religionmonograph. BLIER, SUZANNE PRESTON. I987. The anatomyof architecture: i66 CURRENT ANTHROPOLOGY Volume 32, Number 2, April I99I

Ontologyand metaphorin Batammalibaarchitectural expres- . I957. La methodede l'ethnographie.Paris: Presses Uni- sion. New York:Cambridge University Press. [SPBJ versitairesde France. *i988. MelvilleJ. Herskovits and the artsof ancient Daho- . I96I. Classificationdes insecteschez les Dogon. Journal mey.Res: Anthropologyand Aestheticsi6:I24-42. [SPB] de la Societedes Africanistes3 I:7-7I. .I 989. Field days:Melville J. Herskovits in Dahomey.His- GRIAULE, M., AND G. DIETERLEN. ig5oa. La harpe-luthdes Do- toryin AfricaI6:i-22. [SPB] gon. Journalde la Societedes Africanistes20:209-27. BOUJU, j. I984. Grainede l'homme,enfant du mil. Paris:Societe . ig5ob. Un systeme soudanais de Sirius. Journalde la So- d'Ethnographie. cietedes Africanistes20:273-94. CALAME-GRIAULE, G. I965. Ethnologieet langage:La parole . I95I. Signesgraphiques soudanais. L'Homme: Cahiers chez les Dogon. Paris:Gallimard. d'Ethnologie,de G6ographieet Linguistique3. .I 968. DictionnaireDogon (dialectToro): Langue et civi- . I954. "The Dogon,"in Africanworlds. Edited by D. Forde, lisation.Paris: Klincksieck. pp. 83-I Io. London,New York,Toronto: Oxford University . i987a. Des curis au Marche. Paris: Presses Universitaires Press. de France. . I965. Le renardpale. Vol. I, fasc. I. Le mythecosmogo- . I987b. "Des insecteset des hommes:Essai d'analysese- nique: La creationdu monde.Travaux et M6moires de l'Institut mantiquedes nomsd'insectes Dogon," in Ethnologiques:Hom- d'Ethnologie. mages a Marcel Griaule.Edited by S. de Ganay,pp. 3-I7. Paris: GUERRIER, E. I975. Essais surla cosmogoniedes Dogon: L'arche Hermann. du Nommo.Paris. CLIFFORD, J.I983. "Powerand dialoguein ethnography:Marcel JACKSON, T. I984. "Social anthropological approaches," in Con- Griaule'sinitiation," in Observersobserved. Edited by G. Stock- temporaryapproaches to thestudy of religion. Edited by F. Whal- ing,Jr. Madison: University of Wisconsin Press. ing. Berlin: Mouton. . i988. Thepredicament of culture:Twentieth-century eth- LANE, P. j. I986. Settlement as history:A study of time and space nography,literature, and art.Cambridge: Harvard University among the Dogon of Mali. Ph.D. diss., Universityof Cambridge, Press.[SPB] Cambridge,England. DE GANAY, S. I937. Note sur le cultedu Lebe chez les Dogon du .I988. Tourismand social changeamong the Dogon. African Soudanfranqais. Journal de la Societedes Africanistes7:203-I2. Arts2I(4):66-69. [PJL] .I 94I. Les devisesdes Dogon.Travaux et Memoiresde l'In- LEBEUF, j. P. I987. "MarcelGriaule," in Ethnologiques:Hom- stitutd'Ethnologie 41. mages a Marcel Griaule.Edited by S. de Ganay, pp. xxi-xxviii. . I942. Le binouyebene. Miscellanea African Lebaudy 2. Paris: Hermann. [SPBJ DE HEUSCH, L. I985. Sacrificein Africa:A structuralistapproach. LEIRIS, MICHEL. I934. LAfriquefant6me. Paris. Bloomington:Indiana University Press. .I948. La languesecrete des Dogon de Sanga (Soudanfran- 1i987. Surles tracesdu renardpale. Film,Centre AV a cais). Travaux et Memoires de l'Institut d'Ethnologie so. Bruxelles. LETTENS, DIRK. I97I. Mystagogieet mystification:Evaluation DIETERLEN, G. I94I. Les imes des Dogon. Travauxet Memoires de l'oeuvrede Marcel Griaule.Bujumbara: Presses Lavigerie. de l'Institutd'Ethnologie 40. MICHEL-JONES, F. I978. Retourau Dogon: Figuredu double et .I 959. Mytheet organisationsociale en Afriqueoccidentale. ambivalence.Paris: Le Sycomore. Journalde la Societedes Africanistes29:II9-38. N'DIAYE, F. I972. Contributiona l'etudede l'architecturedu pays *i962. Note sur le totemisme Dogon. L'Homme2:I06-IO. dogon. Objets et Mondes i2:269-86. I967. Colloque surles culturesvoltaiques (Sonchamp, 6-8 OVENDEN, M. W. I976. Mustard seed of mystery.Nature 26i: decembre i965). Recherches Voltaiques, no. 8, pp. 33-44. 6I7-I8. .i982. Le titredhonneur des Arou (Dogon-Mali).Paris: PALAU-MARTI, M. I957. Les Dogon. Paris: Presses Universitaires Musee de l'Homme/Societedes Africanistes. de France. 1i989. Masks and mythologyamong the Dogon. African PAQUES, VIVIANA. I964. L'arbrecosmique dans la penseepopu- Arts23(3):34-43. laireet dansla vie quotidiennedu Nord-Ouestafricain. Travaux DIETERLEN, G., AND J. ROUCH. I97I. Les fetes soixantenaires et Memoiresde l'Institutd'Ethnologie 70. chez les Dogon. Africa4I:I-II. PARIN, P., F. MORGENTHALER, AND G. PARIN-MATTHEY. DOUGLAS, M. I967. If the Dogon.... CahiersEtudes Africaines I963. Die Weissendenken zuviel: Untersuchungenbei den Do- 7:659-72. gonin West-Afrika.Zurich: Atlantis-Verlag. .I968. Dogon culture: Profaneand arcane. Africa38:I6-24. PAULME, D. I937. La divinationpar les chacals chez les Dogon de FORDE, D. Editor.I954. Africanworlds. London, New York,To- Sanga. Journalde la Societedes Africanistes7:I-44. ronto:Oxford University Press. . I939. Parente a plaisanterieet alliancepar le sangen GARTRELL, B. I979. Is ethnographypossible? A critique of an Afrique occidentale. AfricaI2:433-44. Africanodyssey. Journal of Anthropological Research . I940. Organisationsociale des Dogon (Soudanfrancais). 35:426-46. Paris: Editions Domat-Montchrestien. GOODY, JACK. I967. Reviewof: Conversationswith Ogotom- .I948. Condition of women in two West Africansocieties: mei, by Marcel Griaule (London: OxfordUniversity Press, I967). Dogon, French Sudan, Kissi, French Guiana. Africai8:302. AmericanAnthropologist 69:239-4I. .I986. Le statuedu commandeur.Paris: Plon. .1987. Theinterface between the written and theoral. Cam- PAULME, D., AND D. LIFCHITZ. I936. La fete des semailles en bridge: Cambridge University Press. I935 chez les Dogon de Sanga. Journalde la Societedes Afri- GRIAULE, M. I933. La missionDakar-Djibouti. L'Afrique Fran- canistes6:95-II0. ,aise, May, pp. 283-85. PELTON, R. D. I980. The tricksterin WestAfrica: A studyof *I934. Les flambeursd'hommes. Paris. [SPBJ mythicirony and sacreddelight. Berkeley: University of Califor- I935. Burners ofmen. Philadelphia.[SPBJ nia Press. .I938. Masques dogons.Travaux et Memoiresde l'Institut PERN, S., B. ALEXANDER, AND W. E. A. VAN BEEK. i982. d'Ethnologie 33. Masked dancersof WestAfrica: The Dogon.Amsterdam: Time- .I940. Remarquessur le mecanismedu sacrificedogon. Life Books. Journalde la Societedes AfricanistesIO:I27-29. PESCH, P., AND R. PESCH. I977. The Dogon and Sirius. The Ob- *I945. Les grandsexplorateurs. Paris. [SPBJ servatory97:26-28. I948. Dieu d'eau: Entretiensavec Ogotommeli.Paris: Edi- ROTH, P. A. I989. Ethnographywithout tears. CURRENT ANTHRO- tionsdu Chene. POLOGY 30:555-6I. . i952. Les savoirsdes dogon.Journal de la Societedes Afrn- ROUCH, J. I978. The politicsof visual anthropology[interview by canistes22: 27-42. D. Georgakaset al.l. Cineaste8(4). [PIC] VAN BEEK Dogon Restudied I I67

SACCONE, j. I984. Dogon del Mali e le ricerchidel Marcel Gri- . I988. Functionsof sculpturein Dogon religion.African aule. Bologna. Arts2I(4):58-66. SAGAN, C. I979. Broca's brain:The romanceof science. London: . iggoa. "Processesand limitationsof Dogon agricultural Hodderand Stoughton. knowledge," in The growth of ignorance. Edited by M. Hobart. SPERBER, D. I975. Rethinkingsymbolism. Cambridge: Cam- London:School of Orientaland AfricanStudies. bridgeUniversity Press. . iggob. "The sorcerer'sapprentice: Coping with evil in two STRATHERN, M. I987. Out of context:The persuasivefictions of Africansocieties, Kapsiki and Dogon,"in Africanreligions: Ex- anthropology.CURRENT ANTHROPOLOGY 28:25i-8I. [Pic] perienceand expression.Edited by T. Blakeley,D. H. Thompson, TAIT, D. I950. An analyticalcommentary on the social structure and W. E. A. van Beek. London:Currey. ofthe Dogon. Africa 20:I75-99. . I99I. "Enterthe bush: A Dogon maskfestival," in African TEMPLE, R. G. I976. The Siriusmystery. London: Sidgwick and art in the twentieth century: Digesting the West. Edited by S. Jackson. Vogel.In press. VAN BAAL, J., AND W. E. A. VAN BEEK. I985. Symbolsfor com- VAN BEEK, W. E. A., AND P. BANGA. I990. "The Dogon and their munications:Religion in anthropologicaltheory. Assen: Van trees,"in Culturalunderstandings of the environment.Edited Gorcum. byD. Parkinand E. Croll.London: School of Oriental and African VAN BEEK, W. E. A. I983a. Harmonieen schaamtebij de Dogon. Studies. Prana 32:44-53. VAN MAANEN, j. I988. Tales of thefield: On writingethnogra- . I983b. Sacrifice in two African communities. Nederlands phy.Chicago: University of ChicagoPress. TheologischTijdschrift 3:I2I-32. . I987. The Kapsikiof theMandara Hills. ProspectHills: WavelandPress.

Serials

The I989 volume (vol. 7) of AntropologiaPortuguesa 26-27, Quelques finalites pour la rechercheanthropo- consists of the proceedingsof the Second Colloquium logique (50 F); 29-30, L'ethnologueet son terrain,i, Les on AnthropologicalResearch and Teaching in Portugal. aines (70 F); 3I, L'ethnologueet son terrain,2, Les ca- Its contentsinclude "Da antropologiasimb6lica a antro- dets (5o F); 35, L'ethnologiesous contrat(5o F); 36, La pologia cognitiva,"by Ruy Coelho; "Breves consid- recherchesous conditions(5o F); 37-38, La valorisation erac6es sobre o estado da antropologiaem Portugal,"by de la recherche(70 F); 39, Anthropologues,anthropolo- Joaode Pina-Cabral; "Pensamento dogmatico,pensa- gie et musees (60 F). Also available is the publicationof mento positivista: 0 governoletrado das relac6es soci- an internationalcolloquium held in I986 entitled Vers ais," by Raul Iturra;"Funcdo da vizinhanca entreos des societes pluriculturelles:Etudes comparativeset camponeses de Tour6m," by Luis Polanah; "Os Kyaka situation en France (Paris: Editions de l'O.R.S.T.O.M., de Angola," by A. G. Mesquitela Lima; "A questao colo- i987) (i80 F). To join and/orto order,write: A.F.A.- nial na etnologia ultramarina,"by Rui Pereira; "Uma E.H.E.S.S., i, rue du ii Novembre,92I20 Mont- hist6ria da antropologia;Reflex6es sobre a sua necessi- rouge,France. dade e as suas condic6es," by JacquesHouart; "Arte primitiva/artemoderna, encontros e desencontros,"by Ethnologiede la France, a monographseries produced Jos6Ant6nio B. FernandesDias; "Utilizacdo de fontes by the Mission du PatrimoineEthnologique of the Mini- hist6ricasna investigacdoantropol6gica: 0 case de Indo- sterede la Culture et de la Communication,includes as n6sia," by Maria JohannaC. S. Patuleia; "T6cnicas cor- its most recentnumber Bernadette Lizet's La bete noire: porais na arte da danca," by Maria Jos6Fazenda; "Elec- A la recherchedu cheval parfait(i989), an ethnohistoric tra:Uma abordagemantropol6gica da trag6diaatica," by studyof the black horse of Nievre. Among otherrecent Ana Isabel Rodrigues;and "Que ensino de biologia na issues are Ni vue ni connue: Approcheethnographique licenciaturaem antropologia?"by M. L. Rodriguesde de la culturebourgeoise (I988), by BeatrixLe Wita; Le Areia. For furtherinformation about this series and re- sang et la chair: Les abbatoirs des pays de l'Adour lated publications,write: Museu e Laborat6rioAntropo- (i987), by Noelie Vialles; Ethnologiesen miroir:La l6gico, Institutode Antropologia,Universidade de France etles pays de langue allemande (i987), editedby Coimbra, Apartado 3049, 3049 Coimbra Codex, Por- Isac Chiva and Utz Jeggle;and L'herbe qui renouvelle: tugal. Un aspect de la medecine traditionnelleen Haute- Provence (i986), by PierreLieutaghi. Linked to the Subscriptionto the Bulletin de l'Association Frangaise monographseries is a series of Cahiers containingcon- des Anthropologuesand to the association's newsletter ferenceproceedings and the like; topics of recentnum- Les nouvelles breves is a functionof membership(an- bers include the ethnologyof contemporaryurban and nual dues i8o F-8o F forstudents). Back issues of the industrialsocieties, habitat and space in the ruralworld, Bulletin still available include the following:i8, Les and culturesof work. Write: Centre Interinstitutionnel publications (35 F); 20, Rechercheet/ou developpement pour la Diffusionde publicationsen sciences humaines, (30 F); 25, Vers des societes pluriculturelles,2 (20 F); I3I, bd. Saint-Michel,75005 Paris, France.