Navigation – Plan du site
études et essais

Griaule's Legacy: Rethinking "la parole claire" in Dogon Studies

Andrew Apter
p. 95-129


L'héritage de Griaule : repenser la Parole claire dans les études dogon.
Comment lire l'œuvre de Griaule et évaluer son héritage ethnographique ? Que peut-on tirer de son projet ethnophilosophique ? Pour répondre à ces questions, je propose une relecture critique de son ethnographie dogon ainsi qu'un nouveau modèle du savoir ésotérique qu'il prétendit révéler. Ma relecture s'appuie sur deux approches méthodologiques qui redéfinissent le projet exégétique en des termes socialement plus dynamiques. La première approche, qui se fonde sur ma recherche sur les Yoruba au Nigeria, s'appuie sur l'hypothèse selon laquelle les niveaux ésotériques des systèmes philosophiques africains sont en réalité indéterminés et instables, et cette capacité à contredire ou bouleverser le savoir officiel ou ésotérique rend le savoir secret opératoire et donc puissant. La seconde approche méthodologique fait passer l'analyse élaborée de la langue et du symbolisme dogon de l'École de Griaule au niveau de l'analyse pragmatique, plaçant les symboles et schémas dominants ainsi que les genres discursifs rituels dans leurs contextes perfomatifs. En nous concentrant sur les actes de langage, les locatifs, le glissement pronominal et les idées qu'ont les Dogon de leur mise en scène linguistique, nous pouvons revenir sur les riches matériaux dogon et en tirer un modèle dynamique de l'action critique comme l'héritage durable de « la parole claire ».

Haut de page

Notes de l'auteur

An earlier version of this article was presented at Jean-Loup Amselle's seminar on "L'élaboration des savoirs" at the École des hautes études en sciences sociales (EHESS) on 11 January 2001.   I am grateful for valuable feedback from Professor Amselle and his students, and for an extended interview so graciously granted by Professor Jean Jamin during my visit.   I also thank Ralph Austen for insightful comments on an earlier draft.   All errors and opinions are of course mine alone.

Texte intégral

  • 1 If Griaule's favored fieldwork metaphor was that of "juge d'instruction" (JAMIN 1982b: 87; CLIFFORD (...)
  • 2 For a critical reappraisal of Hountondji's position, see APTER (1992b).
  • 3 For more specialized research and discussion of Griaule and his ethnography, see AMSELLE (2000); BO (...)

1Few Africanists, indeed few anthropologists, remain as controversial as Marcel Griaule.   Hailed as a hero of French Africanist ethnography, with a prodigious output and prestigious "school", and assailed as an anti-hero whose sympathy for Africa masked deeper forms of colonial violence, Griaule embodies the best and the worst of our disciplinary history.   Lecturing in his aviator's uniform, badgering his informants1, seizing the Parisian limelight, using aerial photography, embarking on missions, collecting for museums, and eventually "becoming" a Dogon elder, Griaule remained a curious combination of nineteenth century adventurer and 20th century colonial commandant, at once an agent of the French government and a liberal advocate of African cultural sophistication.   Through the extended critique of Lettens (1971), the signal essays of Jamin (1982b) and Clifford (1988), and Hountondji's attack against ethnophilosophy (1983), Griaule became a favorite target, personifying the violence and duplicity of colonial ethnography and its mystification of cultural traditions2.   With his substantive research on Dogon deep knowledge questioned on empirical grounds (van Beek 1991a), his interpretive focus on secrecy and hidden meaning—what he called "la parole claire"—has become iconic of the colonial imagination at large (Mudimbe 1988).   How, then, do we read Griaule's œuvre, and assess its ethnographic legacy?   What can we gain from his ethnophilosophical project?   To answer these questions, I propose a critical re-reading of his Dogon ethnography and a new model of the esoteric knowledge that he purported to reveal3.   My aim is neither to defend nor dismiss Griaule's ethnography on political grounds, but to grasp the inner connections it manifests between language, secrecy, and agency.

2My re-reading is based on two methodological moves that recast Griaule's exegetical project in more socially dynamic terms.   The first move, based on my Yoruba research in Nigeria, is that esoteric levels of African philosophical systems are actually indeterminate and unstable, and that this capacity to contradict or subvert official or exoteric knowledge renders secret knowledge transformative and thus powerful.   Following Griaule, we can acknowledge the significance of what he called "la parole claire"—the deepest level of secret knowledge—as an important domain of social knowledge and power, but contra Griaule, we maintain that its content is fluid rather than fixed.   As we shall see, this perspective vitiates van Beek's "discovery" that the Dogon today do not recognize the secret myths and cosmogony documented in Griaule's later work, since we need not presume that deep knowledge possessed a fixed content in the first place.   If Dogon esoterica is like Yoruba deep knowledge, it is context-specific, not stable and timeless.

3The second methodological move, developing out of the first, shifts the Griaule school's elaborate analysis of Dogon language and symbolism—and more importantly, of Dogon ideas about language and symbolism—to the level of pragmatic analysis, locating dominant symbols, schemas and ritual speech-genres in their contexts of production.   Focusing on speech-acts, locatives, and pronominal shifting, as well as on Dogon ideas about linguistic performance, we can return to the rich Dogon material in terms of its situated pragmatic functions.   In so doing, the central figure of the body in Dogon symbolic classification emerges within an interactive framework of linguistic practice, what Hanks (1990, 1992) calls a "corporeal field" establishing "the indexical ground of deictic reference".

4I will begin by returning to Griaule's "initiation" into the realm of Dogon esoterica, for it is here that his "deep knowledge" paradigm is introduced and developed through conversations with Ogotemmêli, the blind Dogon sage.   It is not my intention to endorse Griaule's self-proclaimed induction, which has become a textbook case of ethnographic mystification, but to identify the model of knowledge he proposes and relate it to my Yoruba material.   From here I will turn to Leiris' formidable study, La langue secrète des Dogon de Sanga, focusing on the pragmatic dimensions of Dogon ritual language that his research brought to light, and which also appear in the ethnolinguistic investigations of Marcel's daughter, Geneviève Calame-Griaule (1986).   I will then extend this communicative framework to Griaule's symbolic codifications, arguing that the dominant Dogon cosmological configurations are better seen as generative schemes for orienting the body in social space, coextensively with broader domains of the habitus.   How these configurations developed dialogically, and in what senses they could be politically transformative, involve fresh considerations of political context and agency that Griaule himself so assiduously repressed.

La Parole claire

  • 4 In addition to GRIAULE (1965, 1983), see also GRIAULE & DIETERLEN (1951, 1991); CALAME-GRIAULE (198 (...)

5The famous Conversations with Ogotemmêli, first published in 1948, represents a turning point in Griaule's research, away from the objectivist documentation of his earlier Masques Dogons (1983) and into the rarefied domain of secret meaning and knowledge.   After fifteen years of energetic inquiry into the manifold activities of Dogon custom and culture, Griaule was "initiated" into the inner sanctum of what he would call "la parole claire", the deepest level of knowledge.   Or so the story goes in its various versions.   Unbeknownst to him, the Dogon elders of Ogol and Sanga held a special meeting and decided to reveal the mysteries of their religion. Griaule was summoned to Ogotemmêli's inner sanctum, where he would return for the next thirty-three days to receive instruction, acquiring the interpretive keys to Dogon culture and society through the wisdom of the ancestors.   From that point on, the Griaule "school" became a scholarly cult, investigating the world of words, signs, and mythopoetic correspondences throughout the Dogon and Bambara regions of the French Sudan4.

6However contrived this famous mise en scène in establishing his ethnographic authority, Griaule was onto something interesting.   Whatever the status of these privileged conversations, and the dialogical text that has come to represent them, Griaule sketches a system of symbolic connections represented by a stratified series of restricted "words."   Never mind that the three words or levels in Conversations, condensed in his shorter essay Descente du troisième verbe (1996), would later be extended to four (Griaule 1952), or that the ordinal series in other contexts collapsed into a binary "simple" vs. "deep" contrast, recapitulated as "exoteric" vs. "esoteric" knowledge, or as paroles de face vs. la parole claire.   Whatever fictions he deployed or illusions he held in his quest for the secrets of Dogon cosmology, Griaule clearly demonstrated the importance of language as its central organizing principle.   Indeed, Geneviève Calame-Griaule (1996: 14-15) would later note the Dogon predilection for reflecting on language, proclaiming that through Ogotemmêli "Dogon civilization revealed itself as a civilization of the Word".   Where father and daughter went wrong, I shall argue, was not in their linguistic formulation of deep knowledge, but in their approach to its meaning and content.

7Griaule was convinced that la parole claire represented a highly restricted and specialized body of knowledge, consisting of myths, codes, signs and classification systems; that is, of philosophical ideas, symbolic associations, and ritual techniques and procedures that together formed a sophisticated ethnophilosophy, one that structured and ordered the Dogon world. As his classic Conversations reveals, together with his collaborative research with Germaine Dieterlen, this knowledge took the form of creation myths—including those of the Nommo Twins, the cosmic egg, and of the pale fox or divinatory jackal—and their symbolic analogues in the material world, particularly granaries, homesteads, villages and fields, all variably mapped onto the human body.   Part Cartesian rationalism, as with Dogon zoological classifications, and part symbolist poetics, resonating throughout social and celestial domains, this knowledge was compared with ancient Greek philosophy, as a precolonial African tradition that was equally rich and worthy of investigation.   Indeed, for Griaule and his intimate group, the pursuit of la parole claire became an end in itself, an almost exegetical mise en abîme that became increasingly unreadable the deeper it went.   But however obscure its myths, graphic signs, and secret languages, such deep knowledge was for Griaule a coherent tradition, a fixed corpus of hidden insights and connections shared by initiated elders of the highest order. Griaule actually estimated the distribution of such specialized knowledge numerically, ranging between 4-15% of the population depending on the relative "word" or level of depth.   Bracketing the question of their quantitative accuracy, these figures framed la parole claire in the form of empirical documentation.

8It is just such an account that van Beek (1991a) has disputed, subjecting Griaule's entire corpus of ethnophilosophical investigations—the period following his putative "initiation"—to sustained empirical and theoretical critique.   So devastating is van Beek's attack, a deconstruction in the literal sense of the term, that it warrants reconsideration to see if any of Griaule's paroles survive.

9Van Beek's essay made quite a splash in the pages of Current Anthropology, where, following the journal's format, it appeared with responses from a range of prominent scholars to whom he then replied.   Presented as a restudy, it offers a "field evaluation" of Griaule's Dogon ethnography, raising a host of interesting questions about secrecy, method, and the reproducibility of ethnographic findings.   After an admirable synopsis of Griaule's two "initiatory" texts—Dieu d'eau and Le Renard pâle—van Beek seeks to demonstrate that most of the deep knowledge elicited by Griaule was actually a "hybrid" product of his creative and fertile imagination, generated in dialogue with a few idiosyncratic informants who did not represent Dogon culture at large.   Attending to what Pels (1994) calls the préterrain of the ethnographic situation, van Beek convincingly argues that the Dogon's "courtesy bias" toward powerful figures of colonial authority, as Griaule clearly was with his pith helmet, whites, and indomitable style, would have lead them to endorse what he wanted to find.   For example, van Beek's discussion of how Griaule neglected the tonal system in his etymologies (van Beek 1991a: 151-2), and how he badgered his informants into producing ad hoc names for twenty-four types of dung beetle (ibid.: 154) without appreciating the inventiveness involved, reveals systematic distorting mechanisms (if not shades of Kafka!) at work.   Equally misleading was Griaule's commitment to a pristine precolonial model, one which blinded him to the way Dogon culture incorporated the foreign within its horizons—a process including ideas, values and techniques of neighboring peoples but also extending to biblical and Q'uranic episodes at the center of his esoteric creation myths.   Indeed, as van Beek points out, Sanga was an important market and administrative center, "and the earliest Christian and Muslim influence radiated from it" (ibid.: 143).   These and other examples of Griaule's interpretive excesses and limitations are indeed important and well established, but the bulk of van Beek's negative commentary comes from his literal-minded understanding of what deep knowledge should be.

  • 5 See BOUJU (1991) for corroboration on this point.

10Van Beek's most devastating evidence against Griaule is lack of corroboration.   When confronted with the contents and contours of Griaule's esoteric writings, the Dogon whom van Beek interviewed were bewildered: "The Dogon know no proper creation myth; neither the version of Ogotemmêli nor that of Le Renard pâle is recognizable to informants", adding that "Dogon society has no initiatory secrets beyond the complete mastery of publicly known texts" (ibid.: 148).   More important than Griaule's misleading emphases on ancestors, the Nommo spirit, and classificatory schemes is the total absence of la parole claire within a socially restricted set of ideas.   Van Beek found that the concept of nyama, or "vital force" is irrelevant to Dogon religion; that body symbolism is not isomorphic with house plans, fields or villages; and that "no sign systems or hierarchical ordering of different paroles (so) or levels of knowledge can be found" (ibid.: 148).   What secret knowledge there is, he argues, pertains to witchcraft, sorcery, and those skeletons in the closet that cast shame upon specific persons and groups whose unbridled ambitions generate conflict and rivalry to undermine the public good5.   As he concludes, "The tendency towards the creation of increasingly ‘deep knowledge' shows itself much more towards the end of Griaule's life, with a decreasing amount of ‘Dogon-ness' marking the text" (ibid.: 157).   Thus the deeper we go the more we get of Griaule, and the less of the Dogon themselves.

11Van Beek's findings were embraced by Griaule's critics and detractors, and provide an important corrective to his lack of historicism and sociopolitical grounding, but I will argue that the negative thrust of the article is misguided.   Contra his claim that Griaule's Dogon present a "paradigm anomaly" in the regional ethnography, owing to the lack of comparable deep knowledge in West African societies (ibid.: 142), a glance at the relevant literature shows otherwise.   Many of Griaule's original team did provide comparable evidence among Bambara, Bozo and Mande peoples, as did Leiris among the Dogon as well.   Certainly research on African ritual associations and secret societies reveals a broad distribution of such "gnostic inner circles of knowledge", as Douglas (1991: 162) points out in her critical response.   It is precisely because Griaule's deep knowledge formulations resonate with my own Yoruba research that I have continued to take them seriously—not in terms of a dogma to be mastered but as a rhetorical resource to be deployed.

12When I began my research on Yoruba religion and politics, or what I now call the hermeneutics of power, I immediately ran into serious barriers.   My project seemed straightforward enough—to examine the politics of orisha worship as a space of contesting political authority—but even my most preliminary inquiries were blocked from the start.   In the kingdom of Ayede-Ekiti, where I conducted my fieldwork, I was allowed to record and photograph public festivals and visit the dominant town shrines, but discussion of ritual symbols and practices, and virtually anything relating to the priests and priestesses, was extremely limited at best.   Over the months and years, I built up enduring relations with ritual specialists and devotees of Yemoja, Orisha Ojuna, and Orisha Iyagba as well as with members of the Ogboni secret society, but I never once experienced any running exegesis of sacred symbols and their secret meanings.   Direct questioning would provoke such responses as "Asáa ni" ("it is tradition"), "Mi m" ("I don't know"), and only eventually "Awó ni" ("it is a secret"), as if the very acknowledgment that there was a hidden meaning was a major concession to my intrusive requests.   Even when my interviews were prepared with schnapps and kola nuts, further demands for money often followed from those who did not know me well, based on the understanding that deep knowledge, called îmo îjinlè, was restricted, scarce, and had extraordinary value.   If I protested such demands—and they could be quite exorbitant—I was asked whether or not specialized education was free in America, and furthermore, if I had any idea what initiates invested over the years to acquire their secrets.   Why would I presume that it should be given freely to me?   Didn't I know that if its secrets were leaked, an orisha would lose its power?

  • 6 This methodological insight was already developed by JAMIN (1977).

13While invoking their orisha during sacrifices and festivals, devotees often pray not to leak any secrets, requesting ritual assistance in sealing their lips.   An elder devotee is much like a vessel, filled with the omi or "water" of the orisha's power which must not leak, spill, or fall—mirroring the vessels of tun, or revitalizing water that they carry on their heads to ritually reproduce the kingdom.   When I began my research on Yoruba ritual and politics, a neighboring chief with a university degree laughed as he told me: "No matter how much time you spend with them, the priestesses will never tell you their secrets—they will never reveal anything!" And indeed, in terms of overt disclosures, he was right.   Formal initiation would not solve the problem since it was predicated on a blood-oath (îmùlè) not to reveal cult knowledge to outsiders.   Indeed, secrets were equally protected within the shrines, distinguishing degrees and grades of elderhood within.   If I were formally initiated, my lips too would be sealed.   Pondering my predicament with the help of palm wine, I experienced something of an ethnographic epiphany.   The very barriers that so effectively blocked my access to deep knowledge should not be seen a problem to overcome, but were themselves part of the ethnographic solution, to be documented as socially significant data.   If the secrets remained forever out of reach, the mechanisms protecting them were not6.

14With my research perspective thus radically readjusted, my fieldwork took a productive turn.   I no longer asked inappropriate questions unless seeking the limits of discursive disclosure.   As I spent more time with the devotees, and was brought into the protected groves and inner chambers of their shrines, I came to appreciate how secrecy operated in practice, as a mode of drawing boundaries, setting agendas, and discussing controversial affairs.   From this method of prolonged osmosis—very different from Griaule's concentrated instruction—I received insights about deep knowledge in fragments, with respect to prior conflicts in the kingdom, in relation to specialized passwords and handshakes, regarding witchcraft and fertility, or pertaining to "true" histories (itàn) not publicly acknowledged, and sometimes danced rather than spoken.   In the more general terms of everyday use by uninitiated Yoruba townsfolk, deep knowledge was associated with powerful people with access to the original secrets of the first ancestors—incantations like àyàjó that were uttered by the first people, true histories proclaimed in the heteroglossic arenas of îjálà chants, in curses (èpè) and incantations (of) that could kill, and invocations of ancestral spirits of humans and deities alike, as in oríkî and ègún pípè, not to mention the enigmatic secrets (awo) of Ifa divination.   Applied to issues or events, however, the qualifier "deep" denoted political disruption and moral violation, as in the usurpation of a royal dynasty by the Eshubiyi line, the murder of a kinsmen for making money-magic (Apter 1998: 84-86), or the nefarious activities of the once ruling National Party of Nigeria (NPN) which rigged gubernatorial election results in 1983.

15Two aspects of such secrets are especially relevant to van Beek's reassessment of Griaule.   The first concerns their subversive character as icons and indices of sociopolitical revolt.   The dominant symbol of the royal Yemoja festival in Ayede, for example, is the calabash (igbá) of concentrated ritual potency se) that is carried—balanced on the high priestess's head—from the bush to the palace, where it empowers the king's person and revitalizes the body politic.   Like any dominant symbol, it embraces a span of meanings ranging from explicit normative blessings ("it brings children and wealth, it keeps the king healthy") to implicit, forbidden themes of division and bloodshed, and it is this latter pole that is powerful and deep.   Yemoja's fructifying calabash represents the womb of motherhood, the head of good destiny, the crown of the king, the integrity of the town, even the cosmological closure of sky and earth.   But its surfaces are decorated with signs of a deadlier power within, indicated by red parrot feathers (îkó odídé)—signs of ritual negation.   Evoking the witchcraft of the priestesses and their mechanism for deposing the king, red parrot feathers on the calabash simultaneously assert a broken womb, miscarried delivery, bad destiny, a decapitated (and crownless) king, as well as political fission and a cosmos out of control.   Such negative themes are rarely voiced in public, but they nonetheless constitute a repertoire of potential interpretations that under certain conditions can be invoked to mobilize opposition against the status quo.   The deep knowledge of royal ritual actually involves the king's sacrifice and rebirth, whereby his icons of personal power and royal authority are literally taken apart and reassembled by authorized priests and priestesses, culminating in the crowning moment, as he receives the calabash, when the king is recapitated, reinstalled and reproduced.   In the case of non-royal festivals, the orisha's calabash serves as a potential crown to remind the king that his chiefs can always rise up and usurp the ruling line.   Such themes are enhanced by various genres of ritual speech, which invoke repressed histories and veiled warnings of former kings and warriors who can prevail again (Apter 1992a: 117-148; 1998).   The dominant visual and verbal tropes that express these themes include those of inversion (e.g. in the image of a capsized canoe), reversal (e.g. from right to left), and mimetic appropriation (e.g. of European crowns).   The latter symbolic function is particularly relevant to bringing outside icons of power within local fields of ritual command, absorbing symbols of foreign value and authority through metaphoric and metonymic associations.

16But if this aspect of deep knowledge invokes fission, usurpation and even militant dismemberment, it does so through mechanisms of formal opposition to received historical and genealogical charters.   The deeper one goes, in a sense, the less fixed and determinate the character of the secret, and the more formal the mechanisms of reversal and inversion.   This second aspect of secrecy as interpretive revision is captured by what I have called the central axiom of Yoruba hermeneutics: "Secret surpasses secret, secret swallows secret completely" ("Awó j'awo lo, awo lè gb'áwo mí torí torí"). Following this axiom to its logical conclusion, the secret behind the secret is that deep knowledge has no content at all but derives its power from context-specific opposition to the authoritative discourses that it implicitly challenges.   Like Griaule's discussion of Dogon esoterica, a salient distinction between exoteric (paroles de face) and esoteric (la parole claire) knowledge is here at play (see also Giaule & Dieterlen 1991: 55), but unlike Griaule, I maintain that the deeper modes have no fixed content into which all ritual elders are eventually initiated.   If the ideology of deep knowledge asserts a fixed corpus of secrets, then this should not be taken at face value, but as a screen that allows its pragmatic functions to masquerade as sanctified wisdom and learning.   As such, deep knowledge is powerful because it is revisionary, sustaining possibilities of political transformation through the revaluation and reversal of established orders.   In the sociocultural contexts of historic Yoruba kingdoms, the political lines of contestation and division are formed around kings in relation to their civil chiefs, and their dominant lines of political segmentation into quarters, lineages and households.

17If we assume, for the sake of argument, that Dogon deep knowledge works in similar ways, not as a fixed corpus of meanings and myths but as an interpretive space of reconfiguration, van Beek's lack of corroborative evidence poses no real threat to la parole claire.   For if it is paradigmatically context-specific, changing over time and space to fit local political groups and relations, we would not expect to "find" the same content at all.   Of course van Beek's critique goes further than this, denying even the formal organization of knowledge into levels of restricted access, but here we can read him against himself.   The deep secrets he did receive, pertaining more to sorcery, witchcraft and skeletons in the closet than to cosmic eggs and primordial foxes, are in fact consistent with the divisive, fissive, and subversive dimensions of deep knowledge as a political—and as we shall see—illocutionary force.   They form part of the repertoire of resources that can be mobilized to challenge the status quo.   And the fragmentary manner in which such data were disclosed is indeed consistent with its mode of dissemination to non-initiates.   I am not implying that van Beek lacked crucial insight into the deeper levels of Dogon symbolism and meaning—his essay on the dama masquerade is a masterpiece of ritual exegesis, linking gender transformations between men and women to the dangerous secrets of social reproduction (van Beek 1991b).   That he does not see such meaning as "deep" is more a consequence of his methodological orientation: if people do not tell him it is deep, or recognize deep knowledge in overt declarations, then in effect it does not exist for him.   I would argue that his Dogon informants could not tell him because they did not know or could not say, given the discursive restrictions against its disclosure.

18To substantiate this counter-critique, and extract the pragmatic kernel of la parole claire from its socially sanctioned mystical shell, I will attend more closely to language itself; first in the more contemporaneous work of Griaule's ambivalent secretary archivist, Michel Leiris, followed by the subsequent investigations of his loyal daughter, Geneviève Calame-Griaule.

Ritual Language

  • 7 That the Vichy regime was uneasy with Leiris is evident when they burned L'Afrique fantôme on the p (...)
  • 8 LEIRIS (1992: 25) even referred to their fieldwork in Sanga as "la mission Lifschitz-Paulme", as if (...)
  • 9 For a poignant account of how the resistance movement originated, to a large degree, in the Musée d (...)

19By the time La langue secrète des Dogon de Sanga was published in 1948, Deborah Lifchitz—of Griaule's Africanist circle—had died in Auschwitz, and it is to her memory that Michel Leiris dedicated this extraordinary study of ritual language.   Published the same year that Griaule's Entretiens first appeared, it presents a complex counterpoint to the French master's voice, at once derivative of his first documentary tome, Masques dogons (1938)—given its extensive focus on sigi ritual texts—and yet uneasy with the colonial situation in Africa and the German occupation of France7.   To what extent La langue secrète voices a muted indictment of Griaule in this broader context is difficult to say, since Leiris' debts cut in both directions: to the 1931 Dakar-Djibouti mission in which he recorded the bulk of his linguistic material, as well as the more "sociological" groundwork first established by Griaule (1938); but also to the "lamented" Deborah Lifchitz, acknowledged in the foreword, with her comrades-in-letters Denise Paulme and André Schaeffner8.   I am not suggesting that Leiris collected his material with a radical political project firmly in place, however haunted he was by the manifold forms of colonial alienation in L'Afrique fantôme; but in Paris, at least, he was writing up in one of the founding cells of the Resistance movement, the basement of the Musée de l'Homme9.   The issue remains unclear, in part, because the politics of secrecy sustains ambiguous interpretations in shifting contexts.

20Ambiguity and unstable content are in fact diagnostic of sigi ritual texts, which are marked more by simplified syntax and morphology than by complex metaphors and poetic associations.   Sigi texts are paradigmatically vacuous, posing difficulties of translation because their shifting meanings so closely relate to their performative contexts.   As Leiris (1992: xv) explains: "[. . .] la nature [. . .] du langage secret, au vocabulaire très réduit et à la grammaire des plus rudimentaires, langue où les choses sont suggérées, indiquées plutôt que décrites, et où la même phrase, la même locution doivent parfois être entendues de façons très différentes suivant le contexte ou suivant les circonstances auxquelles elles s'appliquent, le caractère même de cette langue qui procède sommairement, par larges et brèves allusions, la fait se prêter assez mal à une traduction rigoureuse." These difficulties of capturing ritual meanings highlight the relevance of indexicals in relating texts to their contexts, thereby foregrounding the pragmatics of ritual speech.   The improvisational scope around texts and their variants also suggests contextual sensitivity to fixed liturgical segments, which resonate with implications that extend beyond the meanings of the word themselves.   Indeed, Leiris complains in his preface of the hermeneutical circularity that bedeviled his translations.   Returning to previous sigi transcriptions in order to check or improve their accuracy, his "informateur" would never review "point par point" the same text, but instead would produce a new text—"analogue, certes, mais non pas identique"—as if to underscore the deconstructive joke that every decoding is simultaneously a recoding (ibid.).   Such resistance to definitive philological documentation, however, need not be seen as a problem to be solved, but as a critical characteristic of ritual language itself; that is, as part of the object of description and analysis.   From this more inherently dynamic perspective, suggested to me by the indeterminacy of Yoruba deep knowledge, the goal of achieving definitive translation shifts to that of grasping Dogon interpretations as situated in their social and ritual contexts.   What counts from this more "indigenous" perspective is not an authoritative translation but the plurality of meanings in the public domain.

  • 10 For a discussion of this itinerary over a five-year period, see GRIAULE (1983: 174-175).   See also (...)

21In its broadest sociopolitical context, the secret language of sigi so belongs to an exclusive men's society organized by age-grades and associated with the primordial ancestor Awa, from whom the secrets of sigi descended.   Represented by masked dancers in red fiber skirts, the spirit of Awa embodies the complementary principles of death and rebirth underlying the perpetuation of age-graded generations.   At the apex of the hierarchy is an inner core of olubaru elders, initiated into the highest grade during the full Sigi ceremony, performed every sixty years at the completion of the full generational cycle, and following a coordinated itinerary throughout villages and regions10.   Of course not all Sigi elders would make it to this highest grade, since a man's date of birth determined his point of entry into the overarching cycle.   Hence the marked distinction of this elite body of elders—the oldest of the old—recognized as the highest ritual experts mingling with the ancestors and the dead, and thus referred to as "people of the bush", deemed socially powerful and impure.   These ritual elders were considered the "official depositories" of sigi so, and it was through them that the language was differentially disseminated to the lower grades in accordance with their levels.   Building on Schaeffner's 1935 survey of iron gongs associated with olubaru dignitaries, Leiris estimated their total number as "une quinzaine" in the entire Sanga region, representing a highly restricted inner circle controlling access to secret language and knowledge. Beneath this specialized elite, the remaining elders represent the next level of deep knowledge, acquired through long association with the dama rites and funeral ceremonies connected with the Sigi festival complex.   Following this group are individuals of varying ages who study the sigi language with specialists, short-cutting, as it were, the initiation hierarchy for the conventional fee of 33,003 cowries—a figure signifying a considerable cost rather than a precise amount (Leiris 1992: 17).   The effect of this more direct route to the language was at once financial, providing additional funds for the mask society, but also broadened the social range of ritual language competence and performance.   More conventionally, however, instruction began with male circumcision as the first initiation into the men's society. While prepared and doctored in the bush, the novices received millet beer, sesame oil, black pepper and medicines to help them "hear" the voice of Awa, and began an instruction in sigi language and mythology that increased as they advanced up the hierarchy.

  • 11 For a discussion of different sigi "dialects" in the villages of Banani, Oro Sono, Kunnu and Kabage (...)
  • 12 For extended discussions and studies of the textual status of Mande and Cameroonian epic genres, se (...)

22The content of such esoteric instruction, which increased with elderhood, remains both sketchy and charged in the work of the Griaule school: curiously displaced by each account into shifting mythic figures and linguistic forms.   The initial sigi teachings involve greeting formulae, benedictions and propitiations, myths of origin, the secret names of foods and drinks used in initiation, and a new lexicon for parts of the body.   In addition, initiates learn the exhortations that enhance the nyama, or vital force, of the masked dancers, with specific calls linked to particular masks.   Mythically located in the realm of the bush with ginn (gyinu) spirits and souls of the dead, the sigi language evokes earlier historic migrations by mixing Mande and Voltaic terms and archaisms together with Dogon lexemes11. The mythic origins of the sigi festival itself, and the presiding Grand Awa mask at the center of the ritual complex, trace back to an eponymous ancestor who, after becoming a snake like all old people in that time, broke a taboo against speaking ordinary vernacular to humans, and thereafter ushered in death and mortality, but also regeneration.   Building on Griaule's documentary foundation by providing alternative versions and variants of the origin myth recorded in Masques dogons, Leiris introduced a subtle but significant shift in perspective.   Unlike Griaule, who operated under the philological tradition of a mythic ur-text at the heart of the sigi complex, Leiris recognized the inherent heterogeneity built into the corpus of sigi founding myths.   In an extended footnote to Griaule's "definitive" version, Leiris explains his variations with the insight that "il n'existe pas une tradition unique mais plusieurs traditions, dont il n'y a pas à s'étonner qu'elles soient, sur certains points, contradictoires" (Leiris 1992: 146, n. 2). Through this important interpretive move, Leiris locates contradiction within a corpus of deep knowledge, not as a problem to be eliminated through philological ratiocination but as a feature that makes sense in relation to changing social and performative contexts.   Indeed, his investigations into sigi syntax and morphology highlight the critical relationship between meaning and context12.

23I have already noted that the sigi language is limited and impoverished in comparison with ordinary spoken Dogon, but the direction of its grammatical reductions reveals a significant retreat, as it were, into the body—not as a transcendental symbolic scheme, but as a corporeal field of indexical functions.   Sigi texts are not fixed, Leiris often reminds us, but are rooted in their contexts of ritual production, calling on particular masks, singling out dancers' movements and body parts, saluting presiding spirits, shifting subjects and objects, merging singular with plural, voicing changes of action, shaping points of view, merging past with present, and above all, activating the nyama or efficacy of Awa and his avatars.   To illustrate this dynamic interaction between sigi texts and performative contexts, I will focus on the corporeal dimensions of ritual speech with reference to verbs, pronouns, and locatives.

24Verbs in sigi are few in number, and mainly express movements of dancers or spectators within the ritual arena.   The most frequently used are dyenunu, glossed "to go, enter, or introduce", and sagya, "to place, place oneself, or be placed" (Leiris 1992: 416).   Common utterances exhort dancers to realize and maximize their actions, as in moving those body parts at once protected by the spirits—ostensibly to protect the masked dancer from falling—and activated by being called into motion:

  • 13 See LEIRIS (1992: 198, lines 1, 4, 5, 6, 13- 17).   I have left out the sigi texts and their initia (...)

"Que Dieu garde [vos] jambes!
Que Mouno garde [vos] jambes!
Que Mounokanna garde [vos] jambes!
Que tout le Sigui garde [vos] jambes!
[vos] bras ont remué,
[vos] jambes ont remué,
[vos] yeux ont remué,
[votre] tête a remué,
[votre] corps entier a remué"13.

25As the activation of the eyes suggests, these exhortations shape visual as well as choreographic frames, interposing acts of seeing and being seen within performance arenas.   Verbs are further reduced in terms of tense and mood.   Limited to the present indicative, they represent a condition of "permanent actuality" that for Leiris, citing Leenhardt, quite technically points to the "here and now" (Leiris 1992: 53).   Within this condensed communicative field, only the attitude and tone of the speaker distinguishes imperatives from ordinary indicatives.   In the sigi command "wana boy" ("Venez!"), the urgency with which the addressees must respond—in this case dancers made to get up and perform—is expressed and enhanced by waving the right hand (ibid.: 51).   The dramatic qualities of sigi language, and its enhanced forms of bodily expression, are thus partly a function of its simplified structure.

26Lack of plural and singular forms further reduces the language of sigi to transposable schemas of situated communication.   As in many West African languages, pronouns rather than verbs indicate grammatical number in ordinary Dogon, but in sigi, the pronouns themselves are restricted to the two "floating" forms of first and second persons plural.   The pronominal matrix of ordinary Dogon is thereby reduced to the plural "we" (emme) and "you" (ye/yo/ya), a dyadic opposition that assimilates all singular as well as third person forms, subject only to the quantifier g'ina specifying "all" or "totality".   The effect of such pronominal reduction, combining indifference to time (in the verb system) with indifference to number, is the production of discursive schemas that work, according to Leiris, as a series of dynamic "landmarks" ("repères") bringing fixed relational categories to shape the ritual arenas of movement and participation.   Within this grammar of mythic reduction, plurality and singularity merge through the very acts of exhortation as individual persons, masks and spirits quite literally stand up for groups, categories and types (ibid.: 55-56).   But whereas for Leiris this schematic ordering is iconic of mythic templates and archetypes, we can shift his emphasis from meaning to practice, and discern a transposable interactive scheme.   In context, the pronominal reduction in sigi so to a fixed "I-We/Thou-Ye" opposition assimilates all ritual participants and observers within a dyadic communicative frame.

  • 14 For an analytical formulation of corporeal fields, see HANKS (1990: 91-95).
  • 15 "L'emploi de ces locatifs prête aisément à amphibologie, aucune flexion n'indiquant le rôle joué da (...)

27If the pronominal system merges individual and collective bodies into binary relations of mutual recognition, sigi locatives anchor texts to their contexts by extending the body into a corporeal field—front and back, above and below, inside and outside, right and left, and their associated directional movements (ascending and descending, entering and exiting, etc.) within the performance arena14.   It is thus perhaps no accident that Leiris highlights these terms as particularly confusing and ambiguous precisely because their meanings shift according to their use as substantives, representing parts of the body, or as proper locatives in postposition15.   As he notes, in postposition they indicate spatial relations and only context distinguishes their meanings as particles (ibid.: 57).   The most commonly used locative in sigi is dyu (alt. dyugu) signifying "head" or "body" as a substantive but used in postposition to mean "on", "toward", or "above".   Similarly bin(u) doubles as "stomach" or "in, inside"; poré as "back" and "behind, before" as well as a cadet who is both "behind" in seniority and brings up the rear in ritual procession.   Extending the body to "the various social and personal spaces of an actor" (Hanks 1990: 91), terms like igiru as "earth" or "ground" also signify "below" and the act of descending, contrasting with dara as "sky", "up above", and also used in "ascending".   In a similar contrast, dégu (house, village) and logo (road) signify "inside" and "outside" in more sociocentric terms, and also used in complementary expressions of entering and exiting (Leiris 1992: 421-424).

28These pronominal and locative examples represent those areas of the sigi language most difficult to translate because their meanings are grounded in their performative contexts.   Shifting and "floating" an interactive scheme between phases and components of the ritual process, they reflect a discourse less rooted in Griaule's "mythic substrate" (Ciarcia 2001: 218) than in transformations of person and place.   Sigi so does invoke specialized mythic knowledge in fragments, but does not semantically contain it.   Transcribed and translated, it reveals a privileged scheme of corporeal interaction that mobilizes the nyama of the masked sigi dancers, organizes participants into interposed groups, and directs the flow of movement and energy as the festival unfolds.

29Although the model of linguistic and symbolic analysis remained resolutely exegetical and "cryptological" for the Griaule school, a performative approach to language and to the indexical functions of the corporeal field are clearly adumbrated in Leiris' observations as sketched above, and nearly take shape in the ethnolinguistic investigations of Geneviève Calame-Griaule.   Her Ethnologie et Langage, first published in 1965, builds upon her father's tradition of mythic documentation and symbolic classification in what Dell Hymes acclaimed as a pioneering ethnolinguistic investigation (Hymes 1986: v-vi).   And following her father's focus on the body as an important symbolic template, Calame-Griaule reveals an elaborate corporeal system at the center of Dogon ideas about language, linking speech to the brain, collar bone, liver, heart, pancreas, spleen, larynx and mouth; relating the efficacy of the word to bodily substances like blood and oil; and associating different modes of discourse (greeting, joking, grieving, even teaching) with bodily flows and processes.   Nor is this bodily template restricted to language and speech stricto sensu, but maps onto esoteric graphic signs as the "speech" of various mythic figures and animals.   In this broader context, Dogon ideas about speech and the body provide an interpretive key to the symbolic world at large, and this is the general orientation that Calame-Griaule pursues.   "The word for the Dogon," she writes, "is like a book whose message must be deciphered or ‘decoded', and man is constantly concerned to interpret the ‘signs' around him" (Calame-Griaule 1986: 9). What follows is a marvelous exposition of verbal and symbolic associations, a decoding to be sure, but with surprisingly little structural description or grammatical analysis of Dogon language itself.

30There are revealing moments, however, of pragmatic disclosure that suggest an alternative grounding of her ethnolinguistic corpus, less in the body as an anthropomorphic cipher and more as privileged corporeal field, the ground of socially situated discursive agency.   As Calame-Griaule admits: "Yet, in the measure in which every social act implies a verbal interchange, and where each act is in itself a form of self-expression, ‘speaking' is at times synonymous with ‘undertaking' or ‘doing' [. . .] to Dogon thinking, actions and words are linked together, and this is why, in a symbolic sense, one also calls ‘speech' the outcome of an action" (ibid.: 5.).   Her two examples, "his words have gone inside" (s: vmo yoá) and "it has now become tomorrow's speech" (i.e. the work's continuation will be postponed until tomorrow, Né ygo sóy) involve just those shifting locatives (stomach, inside) and deictics (tomorrow) that extend the body in space and time. Characteristically, Calame-Griaule assimilates this performative dimension of language to "a symbolic sense", thereby reducing doing to meaning. Nowhere is this limitation more clearly imposed than in her ideas about nàma (nyama) as the mana or animating principle that is activated by speech:

  • 16 Her own, unpaginated citation is to her father (GRIAULE 1983).

"Nàma, the life force, is envisioned as a fluid that circulates through the body carried by blood.   The idea of a nàma has appeared since the earliest Dogon studies where it has been defined as an ‘actualizing energy, impersonal, unconscious, shared by all animals, plants, supernatural beings, all the things of nature' [. . .] the simple act of calling someone by one of his names may produce an increase in his corresponding nàma.   For we shall see that the mechanism of speaking is essentially an action upon the personal life force" (ibid.: 19-20)16.

  • 17 In this respect, the Bambara and Dogon notions of nàma are just like the Yoruba idea of àse.

31In this passage we see how nàma is linked to invoking and calling not merely by a conventional symbolic association but through the pragmatic functions of speech acts.   Elsewhere, we learn that the nàma of a speaker determines the nàma of his speech, and correlates with his power and authority, and ability to command (ibid.: 35)17.   What for Calame-Griaule remains an ideational explanation—words are linked to power and efficacy because of Dogon beliefs about language and the body—becomes from our pragmatic perspective more of a grammatical explanation, based on the very principle of linguistic performativity.   Nàma doubles as a property of discourse and a force of nature because the social relation between both domains in grammatically constituted through illocutionary force.   The body as mediator of this dynamic relation establishes a corporeal field, nothing less than the indexical ground of deictic reference through which social contexts and relations are shaped and remade.

32Fragmentary evidence supports this recasting of the Dogon body as corporeal field.   To begin, the truth-value of utterances is closely associated with bodily position.   As Calame-Griaule explains, "‘True Speech' is uttered by a speaker while sitting down.   The position allows for the harmony of all the faculties: the mind is calm, the water in the collarbone is calm, and words in the same way are controlled and well considered.   Elders gathered under the ‘speech shelter' to converse are always seated, and it is significant that the shelter has such a low ceiling that it would be impossible to stay under it otherwise".   Conversely, speech delivered from standing or walking positions is unstable: "If a person speaks standing up, his words ‘do not go down' (s: sùgoyele), they ‘have no path'.   Such is the case with angry speech or words of the light man: he speaks standing up or walking along, for they have no ‘position', no stopping point (înu sèle)" (ibid.: 63-64).   From this relative contrast between sitting and standing, we begin to glimpse how interactive frames of situated communication structure even the truth value of discourse systemically, in relation to down and up.   We would also need to specify associated coordinates of front and back, left and right, intimated in a more symbolic register by the social dimensions of the naming matrix in which Dogon males are located (ibid.: 472, fig. 41), and which is further "superimposed" on a religious schema of ritual qualities and mystical values (ibid.: 536-538, fig. 43).   The Dogon person is further characterized by a right-left and male-female dualism that lies "at the very core of the self" and emerges "always as an agent of imbalance" despite the best efforts to control it through male and female circumcision (ibid.: 33).   Indeed, "[. . .] we find that the whole person is made up of a material carrier or body distinct from, but in close contact with, the outside world through the constant flow of the four elements. [. . .] It is watered by the nàma fluid, like the land by its streams, and is the source of his vital impulse.   [. . .] Thus the individual is not closed, but open, to the world, soaking it in, it could be said, through every pore".

33I would complement this passive or receptive view of the Dogon person with a grammatically more active voice, extending out from the body as a corporeal field to shape the interactive contexts of individual and collective agency.   Thus when Calame-Griaule asks: "How does the ‘self' make contact with other, similar ‘selves'?   What part of himself does he project outside?   How does he act upon others?" (ibid.: 31-32), we can look not only to the liver, blood and collar bones of the body, but also "the spatial coordinates of bodily orientation" (Hanks 1990: 90) in situated communication.

Rereading Griaule

34In what senses do the pragmatic functions of Dogon language and ritual speech extend to "la parole claire" writ large?   How does nàma (nyama) as an indigenous principle of linguistic performance inform the revisionary power of deep knowledge more generally?   By pursuing these questions, we reanalyze Dogon deep knowledge as a sanctified form of critical agency; one that demands not a rigid approach to secret myths and symbolic codes but a more flexible feel for generative schemes.   In the section that follows, I use Hanks (1990) and Bourdieu (1977) to transpose the body as corporeal field into broader structures and domains of the habitus.   Where Griaule saw the body as both cipher and microcosm of the Dogon world, we treat the body as geometer; that is, as the locus of practical homologies and scheme transfers that generate the "structuring structures" of Dogon society (Bourdieu 1977: 114-124).   By introducing this critical shift in perspective, we can extract the pragmatic kernel from the mystified shell of Griaule's initiatory ethnography, recognizing the importance of those formal homologies and generative schemes that his work brought to light while suspending the mystifying constraint that they hold the same content for an inner circle of elders.   And in developing a more dynamic analysis of "la parole claire" as critical practice, we can further explain why its oppositional features would generate orthodox and heterodox variations—not a fixed body of wisdom but a space of critical dialogue and revision.   This more dialogical approach to the celebrated "conversations" with Ogotemmêli, however, involves a return to political context and agency that Griaule seems to have repressed or ignored.

  • 18 The not so subtle implication of this claim is that the work of his students and others on his team (...)
  • 19 This logic of containment, by which a relation between elements is "counted" as an additional eleme (...)

35Initial evidence of a more indexical approach to "la parole claire" as a body of knowledge comes from a short analytical summary that Griaule published just four years before his premature death in 1956.   Representing his final reflections on the organization of deep knowledge, "Le savoir des Dogon" reveals four degrees or levels of access, ranging from the most general and public to the most esoteric and specialized.   The first level of giri so, which he glosses as "parole de face", provides an abridged digest of the Dogon world in which certain mythic figures are disguised or left out, episodes conflated, and references fairly limited to the realm of the visible.   This preliminary level corresponds to the empirical documentation of his pre-1946 mission, Griaule tells us, representing a first stage in such works as his Masques dogons18.   The second level of benne so, glossed as "parole de côté", comprises words that were left out or forgotten in giri so; that is, meanings that are held back until the person is ready, mainly with reference to Dogon rites and representations.   Here we reach the celebrated turning point of Griaule's Conversations, with its privileged access to secret wisdom revealing a world of hidden insights and symbolic associations.   Beyond, or perhaps we should say beneath, this level, is that of bolo so, or "parole de derrière", which completes the preceding level with still further secrets otherwise withheld, but also synthesizes other domains of knowledge in relation to general, more abstract principles of classification.   In this sense depth and synthetic totality go together as complementary principles of knowledge formation and acquisition.   Represented by his post-1949 writings, the third level of bolo so corresponds qualitatively to the still deeper insights of Le Renard pâle, written with Germaine Dieterlen and published posthumously in 1965, and applies quantitatively to the ethnosemiotic inventory of Signes graphiques soudanais (Griaule & Dieterlen 1951).   And finally, capping, and in a crucial sense, containing, the three levels or "words" of Dogon knowledge is so dayi, or "la parole claire" proper, concerning the "edifice of knowledge in its ordered complexity" but also ultra-restricted insights like the ethnoastronomy of the system of Sirius.   Griaule himself acknowledges a certain flexibility built into this scheme, subject to diverse interpretations.   As he notes, la parole claire is conditioned by various channels of transmission; the family, age-set, occupational group, priesthood, mask society, among others (Griaule 1952: 34).   And as with the language of sigi so, there are regional variations of its distribution among elders.   Even analytically, a tension persists between qualitative form and quantitative content.   La parole claire is at once a comprehensive inventory of secret signs and meanings, and the logical predicate of an ordered triad representing the very principle by which the three preceding words or levels are organized into a framework of knowledge19.   What is more, this very classification of levels remains one of the deepest secrets, safeguarding a hidden logic of relations for the deserving few, after years of perseverance and study.

36Clearly there is a heroic autobiography written into this schematic disclosure, a self-serving synthesis of research projects and fieldwork "missions" that further ratifies the initiatory paradigm and the privileged position of "the European" at its apex.   Equally suspect is the objectivist language of its representation as a consistent and shared body of secrets differentially distributed between 5-6% of upper Sanga at the highest level, and between 15-20% at the penultimate level of restriction.   But one striking feature worth taking seriously concerns the "body" of knowledge as corporeal field.   In the exoteric knowledge of giri so, giri is glossed as "eye, face, in front, straight", extending eye and face of the body into the agent's forward orientation toward other interactants; that is, into social space.   The next level of benne so literally refers to the "side" or "profile", extending the sides of the body and face into left and right axes of the interactive matrix. Within this emerging corporeal field, the third level of bolo so extends backwards from the "derrière" of the paradigmatic social body into a "dorsal" as opposed to the "frontal" axis, synthesizing front to back, as it were, in the acquisition of knowledge.   The fourth level of "la parole claire" performs a logical synthesis by combining the left-right and back-forward axes into a system of corporeal coordinates, thereby framing "the body as it normally engages in movement and action" (Hanks l990: 90).   What is missing from this explicit matrix is the up/down axis of above and below, but we can derive this from the corporeal field as covert dimensions of embodied knowledge that remain embedded in social space and implicit in interaction.

37In what sense, then, does the corporeal model of knowledge extend to the broader social contexts and domains of the habitus?   Through what techniques of the body are corporeal schemes put into practice?   In recasting the Dogon body as geometer, I will develop two related arguments that together account for the power of Dogon deep knowledge and its shifting, revisionary content.   First, I will revisit Griaule's material on homesteads and villages to identify those "generative schemes" and "practical operators" (Bourdieu 1977: 125) serving as "structuring structures" of the habitus. Second, I will examine the political dynamics of Dogon villages and districts, for these charged arenas of competitive action produce the orthodoxies and heterodoxies of deep-knowledge claims.

  • 20 See also PAULME (1940: 314-320).

38If, as Bourdieu (1977: 89) has argued, the house is a "privileged locus" for the objectification and embodiment of generative schemes, then the Dogon homestead should provide a rich source of symbolism for practical reinterpretation.   Like the internal space of the Kabyle house analyzed by Bourdieu (ibid.: 90-92) and the interactive matrix of the Mayan solar discussed by Hanks (1990: 95), the Dogon big house (ginu da) represents a concentrated space of deeply embedded meanings and values.   For Griaule, these meanings are of course mythic and primordial, relating the human form of "Nommo, the Demiurge, the reorganizer of the world" to the house plan with its towers as his limbs; and further relating the heavenly placenta and its earthly counterpart (which together represent the head and legs of man lying on his right side) to the kitchen (his head), stable (legs), central rooms (trunk and belly), store rooms (arms), and entrance (penis).   The logic of the plan is explicitly generative for Griaule, in that the form and structure of the house manifests an on-going extension of original creation.   The sexualized entryway leads by a narrow passage to a workroom containing jars of water and grinding stones; these are used to crush new corn, yielding a liquid associated with male seminal fluid that is carried to the left-hand end of the entry where it is poured on the shrine of the ancestors.   Such libations are a way of "fertilizing" the ancestors, of extending their reproductive powers forward in time.   Griaule explains: "Each part of the building represents an original being germinating and growing from its genitor. The whole plan is contained in an oval which itself represents the great placenta from which have emerged, in course of time, all space, all living beings, and everything in the world" (Griaule & Dieterlen 1954: 97-98)20.

39A practical "reading" of domestic space is structural in the sense that it identifies key oppositions, but following Bourdieu (1977: 90), the meanings are inhabited, "read with the body, in and through the movements and displacements which make the space within which they are enacted as much as they are made by it".   Of central importance to this dialectic of objectification and embodiment is the practical mastery of the fundamental, generative schemes shaping everyday routines—not the symbolic oppositions themselves but their navigation and deployment by the body.   Thus from a practical perspective, inside and outside, empty and full, open and shut, left and right, are significant when activated, as "going in and coming out, filling and emptying, opening and shutting, going leftwards and going rightwards, going westwards and going eastwards, etc." (ibid.: 91).   Like Griaule, we can focus on the body as the organizing scheme of domestic space, but with Bourdieu, it is the body in motion, in its habituating practical routines.

40Returning to the Dogon house with a more dynamic approach to its corporeal oppositions, we enter from the "male" vestibule to the workroom where the jars and grinding stones used to produce "male seminal fluid" for the ancestors are located; turning left we proceed to the central room (dembere) or "room of the belly", moving inside to the female center of the house: "The big central room is the domain and the symbol of the woman; the store-rooms each side are her arms, and the communicating door her sexual parts.   The central room and the store-rooms together represent the woman lying on her back with outstretched arms, the door open and the woman ready for intercourse" (Griaule 1965: 94-95).   Even from this minimal sketch we can grasp the implicit meaning of entering a house —from the outside door of the male sexual organ (ibid.: 94) to the inner "belly" of the central room as an act of male penetration and copulation, corresponding to such oppositions as outside and inside, public and private, active and passive, above and below that are reiterated and semantically saturated throughout the house.   Thus the ceiling is dry and male, associated with departing smoke and measured produce (red pepper, purple sorrel, yellow millet) drying on the roof; and also with ancestors on the carved door and façade of the upper story; the floor inside is moist and female, with jars of water, the place of fertilization and childbirth: "When a child is to be born, the woman in labor is seated on a stool in the middle of the room, her back to the north, and is supported by women.   The infant is delivered on the ground and takes possession of its soul in the place where it was conceived" (ibid.: 95).   Such a fundamental gendered opposition is also transposed—from right to left and left to right, and also along an east-west axis, when man and woman are lying down in bed: "The man lies on his right side facing west, and the woman on her left side facing east.   [. . .] The man lies on his right side and touches the woman with his left hand, never with his right.   The woman sleeps on her left arm, and touches the man with her right.   They never lie in any other position" (ibid.: 95-96). Notwithstanding the large grain of salt with which we should take such archetypal characterizations, clearly a set of qualitative oppositions is associated with this gendered corporeal field, one which also extends back to front, since the hearth is located at the back of the house, and the woman delivers with her back to the north, the entrance and front of the house. Moreover, the impurity of "leftness" is likely linked to the polluting powers of menstrual blood, given its charged symbolism in the men's mask society, when men in red-dyed fiber skirts dance as menstruating women (Paulme 1940: 268-270), and in the strict seclusion of menstruating women for five days in special menstruation huts (ya punune ginu) located at the east-west boundaries of the village (ibid.: 264-268), where they figure as "hands" (Griaule 1965: 97).   And finally, along a vertical axis of gender, we can discern a key opposition that figures in the larger political terms of Upper and Lower Sanga, extending from floor to ceiling.   If, as Ogotemmêli, tells us, "the soil of the ground-floor is the symbol of the earth and of Lébé, restored to life in earth", another reference to the earth as cosmogonic mother, then the face of the house, which "gives its name to the tallest mask", must be none other than the grand mask of Awa, the primordial male ancestor, although the name is not provided in this text.   Ascending is thus associated with Awa, descending with Lébé.

  • 21 Note the male body in GRIAULE & DIETERLEN (1954: 97-98) with female parts, and the female body in G (...)

41A key concept of practical logic is that of scheme transfer, by which generative oppositions like those sketched above are redeployed between contexts, reproducing the same opposition within a wider or narrower social space or field.   The economy of logic, in its practical applications, refers to the interchangeability of a core set of generative schemes over a wide range of social, spatial and semantic domains.   Moving, in a sense, from the inside out, from the core schematic matrix of the household to the broader arenas of the village and district, we can retrace Griaule's interpretive footsteps, recasting what he saw as a widening series of symbolic correspondences into practical homologies extending from the corporeal field. The key coordinates of body space—up/down, left/right, back/front—with their associated domestic values, are thereby transposed into co-extensive political and territorial relations.   For Griaule, these correspondences remained mythic and symbolic: the "body" of the house was reproduced, pars pro toto, in the quarter (or section) of each village, in the village at large, and in the most inclusive district.   As Griaule (1965: 97) explains, "Within the village each quarter is a complete whole, and should be laid out in the same way as the village, like a separate entity". This layout is none other than the anthropomorphic template transposed from micro to macro domains. Like the house, the village (and each of its quarters or sections) replicates the mythic body in its layout and plan, with the iron-smelting smithy at its "head", the family houses as its "chest", the women's menstrual huts as "hands", its central altars as (male and female) genitalia, and its communal altars as feet (See diagram 1).   And like the homestead "correctly sited [. . .] that is to say, open to the north" (ibid.: 94), the body of the village aligns with cardinal points: the head is north, the left and right hands east and west, and the feet south.   In terms of a dynamic corporeal field, a significant egocentric and sociocentric synthesis is thus achieved, whereby north is up, south is down, east is left and west is right.   Because the body is lying down on its back, front is also celestial, back is terrestrial.   Clearly the fixity of such schemes is problematic—as Ogotemmêli tells us, Lower In Griaule's diagram, both male and female genitalia are manifest in complementary altars of the foundational shrine—the male organ in the form of a cone, the female as a hollowed stone.   This dualism replicates the bisexuality of original twinship, with one gender dominating the other, and further suggests that the body of the house and village can be male or female, depending on context21.   But if entering the village is like entering the house, with the male's frontal penetration of interior female space, then we should find maleness outside the village, in marked contrast to female space at the center.   And in fact we do.   In a revealing footnote to the village model, Griaule & Dieterlen (1954: 96, n. 2) explain that "out of respect for the female sex [. . .] and for the women of the village, the male shrine is often built outside the walls" (my emphasis, see also Griaule 1965: 97).   Whether or not this explanation is based on ad hoc secondary elaborations, I see the discrepancy as more consistent with the generative scheme of entering and exiting.   Furthermore, following the logic of scheme transposition, the same opposition can be inverted without contradiction, as context and circumstances dictate.   Thus from the "assembled" perspective of normative politics, of the male regulation of public affairs, men occupy the center in the men's meeting house, while women are marginalized in menstruation huts and the inner rooms of the domestic domain.   In this gendered distribution of male authority and order, with men at the center and women at the margins, a further ethnographic contradiction occurs that vanishes from a practical point of view.   Again, in his normative diagram, Griaule locates the men's meeting house at the "head" of the village, which is above and to the north, whereas in his text, we find such structures in the central square of the village, and in each public square, presumably those of each quarter.   How can it occupy both positions at once—the northern head and the central square?   In this case, Griaule is succumbing to the synoptic illusion (Bourdieu 1977: 97-109), when practical systems of classification are abstracted into a fixed hierarchy of logical relations apart from their context-specific validity, and thus appear contradictory.   The men's house can occupy both positions at once because, according to the coordinates of body space, "up" is male, north, dry, celestial, associated with lineage ancestors and authority, and by implication, "down" is female, south, moist, terrestrial, associated with childbirth.   In effect, the male/female schematism "flips" between north-south, up-down and center-periphery axes, and since Griaule could not diagram the transpositions themselves, he privileged one location at the expense of the other.

DIAGRAM 1. Dogon Village Scheme (from Griaule 1965: 95)

42Ogol is "almost right", and most of the houses of Upper Ogol face west to avoid the prevailing rains (ibid.: 96, 92).   But again I would emphasize practical movement over static design; movement in and out, up and down, back and forth, left and right.   From this more active agentive perspective, we can even "predict" certain amendments to the village scheme that find support in Griaule's text.

43This latter example of scheme transposition shifts into yet another significant generative contrast that relates the words of men to the blood of women.   Returning to the men's house as centrally located in the Dogon village, and "facing north" (Griaule 1965: 98), "the elders [who] [. . .] confer there and take decisions on matters of public interest" (ibid.: 97) are necessarily seated, as required by the corporeal coordinates of "true speech" (Calame-Griaule 1986: 63-64) and structurally imposed by the low roof and rafters. Such true speech, of good words, has a straight path that is cardinally north: as Griaule (1952: 29) reported from native testimony, "la parole de devant et la parole de derrière [sont] bonnes paroles [. . .] car les traits droits connotent une seule parole, ordonnée, droite, parvenant directement à l'auditeur qui doit l'entendre".   In relation to the domestic matrix of gendered orientations, however, these straight-talking elders resemble women.   Like women giving birth within the house, seated on a stool, supported by other women, so that the infant is delivered "on the ground", the men are seated at the center of the town, supported by men, delivering good words that are "grounded" in a ground plan that evokes the seventh and eighth ancestors, "the master of Speech, and [. . .] the Word itself" (Griaule 1965: 98).   But if from the vertical axis of standing and sitting, these men are like women, from the frontal and dorsal axis of ahead and behind, they again assert their difference: women give birth with their backs to the north, whereas men in the speech shelter, at least in principle, face north.   It is the lateral axis of left and right, however, which extends to the most significant contrast between men and women, purity and pollution, center and periphery, and finally, good and bad words.   For at right angles to the north-facing men, the women's menstruation shelters establish left and right "hands" and east-west limits of the village.   So fundamental is this schematic opposition to the construction of gendered public space that "in the past, when a village was founded, the [men's] shelter and the women's houses were the first buildings to be erected" (Griaule 1965: 97).

44And it is in these primary village coordinates that we find the key to the power of la parole claire.   For the secret contrast to the good words of men are the bad words of women as divisive, disorderly, polluted and impure—in short, transfigured as menstrual blood.   According to Ogotemmêli, "the unpleasant smell of the female sexual parts comes from the bad words heard by the ear", adding that "bad words smell" and "affect a man's potency" (ibid.: 139, 142).   Such words are not fertile, but sabotage procreation, since they are formed by unwanted blood, the result of mother-son incest when the jackal (later called the pale fox) violated the earth. Bad words thus pass out of the womb "in emanations" (ibid.: 142).   Good words, by contrast, represent the water of Nummo, which is pure, the female moisture and blood of procreation, mingling with the male seed to create new life.   Since the bad words of menstruating women cause such havoc, they are exiled to the margins of the village, and after their periods, cleanse their vaginas with Lannea acida oil which as a sweetener, works "like a good word in combating the bad smell resulting from a bad word" (ibid.). This profound association of bad words with bad blood explains why we find, at the center of the village, the female shrine as sexual organ, made of stones "on which the fruit of the Lannea acida is crushed" (ibid.: 97), and where the bad blood of women is thus ritually converted into the good words, or strait speech, of men.

45What I am suggesting is that the polluting, transgressive, non-reproductive menstrual blood of women represents the powerful efficacy of deep knowledge, as hidden and secret transformative agency in contrast to the reproductive language of "paroles de face".   Located at the sides and margins of the social body, these "words" have a subversive and destabilizing force as the negative dialectic of cosmological renewal—the valence that sunders and takes apart in order to remake and recombine anew, to purify through ritual reproduction.   In sociopolitical terms, such deep knowledge aligns with dominant sociopolitical cleavages, emphasizing the separation and division of power competition over the administrative unity of the political body as a whole, be it the household, ward, village or district.   The secret of deep knowledge in these oppositional contexts is not found in a hidden doctrine for the privileged few, but in its paradigmatic negation of the status quo associated with the relevant political authority.   If the true words of men are delivered while sitting, evoking the position of women in labor, it is because their official words are socially reproductive.   By contrast, the destabilizing words of the left and right are transformative, false by official standards but sanctified by deeper truth-conditions that remain impure and off-bounds to the public.

  • 22 See TAIT (1950: 176-177) for the full list of villages.

46To support this unorthodox rereading of Griaule, we can sketch the basic contours of Dogon government, focusing on the principal relations of segmentation between competing groups (politics) and their hierarchical inclusion within the region or district (administration).   Studies of Dogon sociopolitical organization reveal a complex tapestry of interrelated and overlapping institutions, including clans and lineages, occupational castes, various cults and mask societies, indigenes and strangers, men's and women's associations, as well as households, villages and districts stratified by age-sets and in many cases divided between Upper and Lower sections with complementary ritual domains.   Thus in the well-worked region of Sanga, divided into Upper and Lower sub-regions, we find villages such as Ogol, Sangui, Ennguel and Bongo, each in turn divided into Upper and Lower sides, as well as villages such Barkou and Gogoli that are not divided as such22.   As Tait's attempted synthesis of the Griaule school literature reveals, there is a lack of consensus on such key points as whether or not the entire Sanga region is lead by one Hogon or two (for each division), how his council of elders and officers are recruited, whether or not the Awa Men's society with its Sigi rites recruits from Upper Sanga exclusively; and if the Lébé cult belongs to Lower Sanga only (Tait 1950).   But such complexities aside, we can abstract a core model of "politics" and "administration" in the Bandiagara cliffs that identifies relations of opposition and inclusion in the following terms: within the district or region, we find two main "halves" (Upper and Lower), segmented into villages, many of which also split into Upper and Lower divisions, and which in turn break down into quarters or wards, and further into lineages and lineage segments organized by households or homesteads of the "grandes maisons".   Headed by the Hogon as a fused political and ritual office, the system was administered through a council of elders that represented the dominant quarters and lineages of each village, and these presumably were similarly organized around a senior elder and his representative age-mates, and so forth in a replicating pattern that applied to the village quarter and even the "houses" within, each with its "chef de maison".   According to this ideal-typical pattern, authority relations refer to the hierarchical inclusion of houses, quarters and villages within the Hogon's regional jurisdiction, and political relations to the divisive competition between houses, quarter, villages, even Upper and Lower divisions—that is, between units of equivalent status and rank—for the power to determine public policy and its outcomes.

47From this structural framework, derived from Smith (1956, 1960), power is transformative and divisive, generating factions from below that attempt to prevail over others, whereas authority is reproductive, containing competitive politics through the execution of public policy from above, as one administrative body.   As Smith points out, this dialectic of power and authority informs all levels of political organization; in the Dogon context, these would include the family and its internal sibling groups up through the ward, village, and district levels.   But what I am suggesting in ritual terms is that the transgressive potency of power sui generis, opposed to authority as it violates its categories and constraints, is simultaneously manifest in menstrual blood and the unbounded powers of the bush, outside of society. Denise Paulme notes that the olubaru elders in the highest grade of the Sigi society are not only considered impure (inne puru) and close to death, but are also immune to menstrual blood and exempt from the taboos that otherwise apply to "hommes vivants".   For this reason they are designated to repair the women's menstruation huts every year, and preside over the sigi and dama masks that are both powerful and impure.   In fact, the red fiber skirts, evoking the original blood of incest and women's subsequent menstrual debt to the earth, transform the powerful masked men of the bush into menstruating women, subject to the same taboos while manifesting their polluting potency:

"Ce caractère impur des masques atteint son maximum le jour où les hommes teignent en rouge les vêtements en fibres de sansevière (jupes, bracelets, etc. [. . .]) qui accompagnent le port des masques.   Anam punia, ‘menstrues des hommes', désigne ce jour de la coloration en rouge des fibres, où les hommes ne se montreront pas plus au village que ne s'y promèneraient des femmes menstruées" (Paulme 1940: 269).

48Paulme further notes that the impure elders of the mask society are charged with the conservation and transmission of its traditions, linking the deepest power (nyama) of the deepest secrets to the polluting blood of women.

49The complementary frame, in political terms, of the "witchcraft" of such unbridled power is the "fertility" of administrative authority; the Hogon in office, the moral order upheld, and the reproductive capacities of men and women revitalized.   In ritual terms, the power of deep knowledge and menstrual blood must be domesticated and incorporated into the body politic, brought from the bush into the center of the village—in effect cleansed and channeled into regeneration.   Effective incorporation within the social order involves a cooling and freshening of the hot redness of Awa, the notion of tawa in sigi so (Leiris 1992: 72) associated with the humidity and wetness of cleansed female sexuality.   Such a reading of the "ardeur-fraîcheur" opposition (ibid.: 73) is consistent with van Beek's interpretation of the dama festival as the male appropriation of female reproductive power, but further identifies menstrual blood with political power and deep knowledge.   We therefore identify the blood of women as a critical component of ritual reproduction—a feature which van Beek cannot really explain—and the domain of deep knowledge not as fixed esoterica (which van Beek could not find) but as a shifting corpus of generative schemes and pragmatic functions.   What makes deep knowledge powerful is not only its association with the bush, but its formal, pragmatic and structural opposition to the authority structures that it enters and revitalizes.

50In the broadest context of Upper and Lower Sanga, the Awa mask associated with sigi and dama clearly mediates political relations between sub-regional levels; between villages themselves (some perform sigi together, others apart); and within villages, between their upper and lower halves, between their quarters, and between lineages.   Clearly the cross-cutting sodalities of age-sets are crucial in this process, dividing opposed sections in political competition and combining them in administrative hierarchy, and it is precisely such relations of union and separation that correspond to freshness and heat, power and authority, outside and inside, deep words and straight speech, values that are contextually situated in the dominant coordinates of the corporeal field.   And clearly, the very politics of village fission and rebellion are signaled by new ritual itineraries and routines.   But to finally explain how la parole claire relates to the semantic and corporeal schemes of the habitus, in the overt symbolic if unstable elaborations which Griaule's ethnophilosophical investigations brought to light, we need to return to the generative conditions of orthodox and heterodox discourses.

51In Bourdieu's initial theory of practice, the world of the habitus—the house, the village, the farming cycle and agricultural rituals—is paradigmatically silent.   Implicitly embedded in social space and time, through the practical taxonomies shaping everyday routines, practical consciousness exists in the body, as a set of habituated dispositions rather than an explicit ideology or body of ideas.   In an oft-quoted statement, "what is essential goes without saying because it comes without saying [. . .] the play of mythico-ritual homologies constitutes a perfectly closed world, each aspect of which is, as it were, a reflection of all the others" (Bourdieu 1977: 167, original emphasis).   In such a field of internalized doxa, there can be no deep knowledge of la parole claire, reflecting on the secrets of empowerment and reproduction, unless either "culture contact" or class formation calls the given world into question.   If Bourdieu has been criticized for hypostasizing such an organic idiom for the traditional "pre-modern" world, or incorporated into a more realistic theory of hegemony (Comaroff & Comaroff 1991: 23-27), his insight stands as a theoretical definition of the embodied forms of practical consciousness.   As such, it remains characteristically uncritical: "The adherence expressed in the doxic relation to the social world is the absolute form of recognition of legitimacy through misrecognition of arbitrariness, since it is unaware of the very question of legitimacy, which arises from competition for legitimacy, and hence from conflict between groups claiming to possess it" (Bourdieu 1977: 168).

52As I have argued in a Yoruba context (Apter 1992a: 228, n. 7), the critical calling into question of doxa—the meanings and values embedded in the habitus—is not limited to "contact" or class formation, but is also found in the dialectics of power and authority motivating segmentary opposition and administrative hierarchy.   Competition for power seeks the revision of authority structures just as administrative authority contains power competition, thus along the dominant political cleavages and lines of segmentation we will find heterodox challenges to the orthodoxy of the status quo.   In the Dogon context, such challenges could be leveled against the chef de maison within a homestead, between lineage elders of a quarter or village, or against the Hogon of Upper Sanga by a hunter or priest of Lower Sanga—perhaps by Ogotemmêli himself.   My point is not to document such challenges, since I can only infer such relations from the data (see also Jolly 1998-1999, 2004; Jolly & Guindo 2003), but to establish the political conditions in which they arise, as a heterodox field of destabilizing discourses against "the straight speech" of men.   Through the ritual articulation of political competition and mediation, the deep knowledge of la parole claire destabilizes the very ground of political authority by reflecting on the generative schemes of the habitus, endowing them with explicit mythic content, and shifting their meanings to remake the body politic.   That such renewal is always transformative and reproductive, hot and cool, polluting and purifying, powerful and authoritative, does not detract from the power of deep knowledge, but locates it within a dynamic arena of competitive politics and material relations.


54Leach (1958: 120) has said that good ethnography sustains alternative analyses and interpretations, including those that controvert the original author.   Thus he praised Malinowski's Trobriand fieldwork while "refuting" his extensionsist kinship thesis in a rival analysis of social category terms. In a sense, I have pursued a similar strategy, affirming the value of Griaule's Dogon studies and the enduring legacy of "la parole claire" in an alternative reading of its indexical functions, privileging pragmatic over semantic categories as shifting coordinates of a corporeal field.   Despite the growing criticism of Griaule's colonial epistemology, his ahistoricism, the violence of his strategies and tactics, and the paternalism of his "sympathetic" quest —problems that belong to anthropology's history more generally—his investigations into Dogon deep knowledge are not so easily swept aside, but provide a body of knowledge, a corpus inscriptionum (to reinvoke Malinowski [1961: 24]), that sustains productive reanalysis and reinterpretation.

55In my own re-reading of selected texts and passages, including those of Leiris and Calame-Griaule, the body of knowledge takes center stage not as a collection of secrets or a symbolic cipher, but as the indexical ground of deictic reference, generating and transforming social space and political context through ritual performance and incantation, and thus challenging and revising authority structures through idioms of cosmological renewal. When secret knowledge is liberated from its ideology of fixed and determinate content, it can be seen as a socially sanctioned rhetorical resource that is unstable and destabilizing because it redeploys the practical homologies of the habitus, drawing on the left hand and the polluting blood of women to reshape the social and political body.   In this capacity, "la parole claire" sustains a grammatical space of critical practice and agency, which if lost on Griaule in his search for hidden symbols, remains equally lost on his critics and detractors.

56UCLA, History Faculty.

DIAGRAM 1. Dogon Village Scheme (from Griaule 1965: 95)
Cara ïbes noirs et negros franceses : des îles au continent

Haut de page


— 2000 "L'anthropologie au deuxième degré: À propos de ‘La mission Griaule à Kangaba (Mali)' de Walter E. A. van Beek et Jan Jansen", Cahiers d'Études africaines 160, XL (4): 775-777.


1992a Black Critics and Kings: The Hermeneutics of Power in Yoruba Society (Chicago: University of Chicago Press).
— 1992b "Que Faire? Reconsidering Inventions of Africa", Critical Inquiry 19 (Autumn): 87-104.
— 1998 "Discourse and its Disclosures: Yoruba Women and the Sanctity of Abuse", Africa 68 (1): 68-97.

— 1995 The Elusive Epic: Performance, Text and History in the Oral Narrative of Jeki la Njambè (Cameroon Coast) (Atlanta: African Studies Association).

AUSTEN, R. A. (ed.)
— 1999 In Search of Sunjata: The Mande Oral Epic as History, Literature and Performance (Bloomington: Indiana University Press).

— 1991a "Dogon Restudied: A Field Evaluation of the Work of Marcel Griaule", Current Anthropology 32 (2): 139-167.
— 1991b "Enter the Bush: A Dogon Mask Festival", in S. VOGEL (ed.), Africa Explores: 20th Century African Art (New York: The Center for African Art): 56-73.

— 2000 "La mission Griaule à Kangaba (Mali)", Cahiers d'Études africaines 158, XL (2): 363-76.

— 1977 The Vildé Affair: The Beginnings of the French Resistance (Boston: Houghton Mifflin Company).

— 1984 Graine de l'homme, enfant du mil (Paris: Société d'ethnographie).
— 1991 "Comments", Current Anthropology 13 (2): 159-60.

— 1977 Outline of a Theory of Practice. Trans. R. Nice (Cambridge: Cambridge University Press).

— 1986 [1965] Words and the Dogon World. Trans. D. LaPin (Philadelphia: Institute for the Study of Human Issues).
— 1996 "Préface", in M. GRIAULE, Descente du troisième Verbe (Paris: Fata Morgana): 11-22.

— 1998 "Ethnologues et ‘Dogon': La fabrication d'un patrimonie ethnographique", Gradhiva: Revue d'Histoire d'Archives de l'Anthropologie 24: 103-115.
— 2001 "Dogons et Dogon. Retours au ‘pays du reel'", L'Homme 157 (janvier/mars): 217-229.

— 1988 [1983] "Power and Dialogue in Ethnography: Marcel Griaule's Initiation", in J. CLIFFORD, The Predicament of Culture: Twentieth-Century Ethnography, Literature and Art (Cambridge, MA: Harvard University Press): 55-91.

— 1991 Of Revelation and Revolution: Christianity, Colonization and Consciousness in South Africa (Chicago-London: University of Chicago Press).

— 1942 Les âmes des Dogon (Paris: Institut d'Ethnologie).
— 1951 Essai sur la religion bambara (Paris: Presses Universitaires de France).
— 1971 "Les cérémonies soixantenaires du Sigui chez les Dogon", Africa 41 (1): 1-11.

— 1999 Les masques dogon: ethnologie savante et ethnologie autochtone (Paris: Karthala).

— 1991 "Comments", Current Anthropology 13 (2): 161-2.

— 1942 Les Devises des Dogon (Paris: Institut d'Ethnologie).

— 1987 "Anatole Lewitzky: De l'ethnologie à la Résistance", Liberté de l'esprit: Visages de la Résistance 16: 237-253.

— 1952 "Le savoir des Dogon", Journal de la Société des Africanistes 22 (1-2): 27-42.
— 1965  [1948] Conversations with Ogotemmêli: An Introduction to Dogon Religious Ideas. Trans. R. Butler (London: Oxford University Press).
— 1973  [1954] "The Mother's Brother in the Western Sudan", in P. ALEXANDRE (ed.), French Perspectives in African Studies (London: Oxford University Press): 11-25.
— 1983  [1938] Masques dogons (Paris: Institut d'ethnologie).
— 1996 Descente du troisième Verbe (Paris: Fata Morgana).

— 1951 "Signes graphiques soudanais", L'Homme 3 (special issue).
— 1954 "The Dogon of the French Sudan", in D. FORDE (ed.), African Worlds: Studies in the Cosmological Ideas and Social Values of African Peoples (London: Oxford University Press): 83-110.
— 1991  [1965] Le Renard pâle (Paris: Institut d'ethnologie).

— 1990 Referential Practice: Language and Lived Space among the Maya (Chicago: University of Chicago Press).
— 1992 "The Indexical Ground of Deictic Reference", in A. DURANTI & C. GOODWIN (eds.), Rethinking Context: Language as an Interactive Phenomenon (Cambridge: Cambridge University Press): 46-76.

— 1983 African Philosophy: Myth and Reality. Trans. H. Evans and J. Rée (London: Hutchinson).

— 1986 "Forward", in G. CALAME-GRIAULE, Words and the Dogon World. Trans. D. LaPin (Philadelphia: Institute for the Study of Human Issues): v-viii.

— 1977 Les lois du silence: essai sur la fonction sociale du secret (Paris: François Maspero).
— 1982a "Les métamorphoses de L'Afrique fantôme", Critique: Revue générale des publications françaises et étrangères XXXVII (418): 200-212.
— 1982b "Objects trouvés des paradis perdus: à propos de la Mission Dakar-Djibouti", in J. HAINARD & R. KAEHR (eds.), Collections passion (Neuchâtel: Musée Ethnographique): 69-100.

— 1998-1999 "Chefs sacrés et chefs de guerre dogon: deux pôles du pouvoir", Clio en Afrique 5: wwww. up. univ-mrs. fr/ ~wclio-af/ numero/ 5/ thematique/ jolly/ index. html
— 2004 Boire avec esprit: bière de mil et société dogon (Paris: Société d'éthnologie).

— 2003 Le pouvoir en miettes. Récits d'intronisation d'un hogon (pays dogon, Mali) (Paris: Classiques africaines).

— 1958 "Concerning Trobriand Clans and the Kinship Category Tabu", in J. GOODY (ed.), The Developmental Cycle in Domestic Groups (Cambridge: Cambridge University Press): 120-145.

— 1987 "Marcel Griaule", in S. DE GANAY et al. (eds.), Ethnologiques: Hommages à Marcel Griaule (Paris: Hermann): 120-145.

— 1992  [1948] La langue secrète des Dogon de Sanga (Paris: Éditions Jean-Michel Place).

— 1987 "L'heure très sévère de Boris Vildé", Liberté de l'esprit: Visages de la Résistance 16: 329-341.

— 1971 Mystagogie et mystification. Évaluation de l'œuvre de Marcel Griaule (Bujumbara: Presses Lavigerie).

— 1961  [1922] Argonauts of the Western Pacific (New York: E. P. Dutton).

— 1999  [1978] Retour aux Dogon. Figures du double et ambivalence (Paris-Montréal: L'Harmattan).

— 1988 The Invention of Africa: Gnosis, Philosophy, and the Order of Knowledge (Bloomington: Indiana University Press).

— 1940 Organisation sociale des Dogon (Soudan français) (Paris: Éditions Domat-Montchrestien).
— 1992 Lettres de Sanga (Paris: Fourbis).

— 1994 "The Construction of Ethnographic Occasions in Late Colonial Uluguru", History and Anthropology 8: 321-356.

— 2000 "La yasigui et Le Renard pâle: mythes, controverses, images...", Cahiers d'Études africaines 159, XL (3): 415-431.

— 1960 La religion et la magie songhay (Paris: Presses Universitaires de France).

— 1956 "On Segmentary Lineage Systems", Journal of the Royal Anthropological Institute 86 (2): 39-80.
— 1960 Government in Zazzau, 1800-1950 (London: Oxford University Press).

— 1950 "An Analytical Commentary on the Social Structure of the Dogon", Africa 20 (3): 175-199.

— 1960 Sociétés d'initiation bambara (Paris: Mouton).
— 1963 La dialectique du verbe chez les Bambara (Paris: Mouton).

Haut de page


1 If Griaule's favored fieldwork metaphor was that of "juge d'instruction" (JAMIN 1982b: 87; CLIFFORD 1988: 74-75), Leiris would write, in L'Afrique fantôme, "Je continue mon travail de pion, de juge d'instruction ou de bureaucrate.   [. . .] Pourquoi l'enquête ethnographique m'a-t-elle fait penser souvent à un interrogatoire de police?" (quoted in JAMIN 1982a: 205).

2 For a critical reappraisal of Hountondji's position, see APTER (1992b).

3 For more specialized research and discussion of Griaule and his ethnography, see AMSELLE (2000); BOUJU (1984); CIARCIA (1998, 2001); DOQUET (1999); JAMIN (1982b); JOLLY (1998-1999, 2004); LEBEUF (1987); LETTENS (1971); MICHEL-JONES (1999), PIAULT (2000); VAN BEEK & JANSEN (2000).

4 In addition to GRIAULE (1965, 1983), see also GRIAULE & DIETERLEN (1951, 1991); CALAME-GRIAULE (1986); DIETERLEN (1942, 1951); GANAY (1942); LEIRIS (1992); PAULME (1940); ROUCH (1960); ZAHAN (1960, 1963).   Despite their self-distancing from Griaule's "school", Leiris and Paulme produced their work through his "missions".

5 See BOUJU (1991) for corroboration on this point.

6 This methodological insight was already developed by JAMIN (1977).

7 That the Vichy regime was uneasy with Leiris is evident when they burned L'Afrique fantôme on the pyre (JAMIN 1982a).

8 LEIRIS (1992: 25) even referred to their fieldwork in Sanga as "la mission Lifschitz-Paulme", as if to further accentuate their differences with Griaule.   See also PAULME (1992).

9 For a poignant account of how the resistance movement originated, to a large degree, in the Musée de l'Homme, see BLUMENSON (1977).   Although Leiris himself "lacked the temperament" to join (Jamin, personal communication), the cell was led by the anthropologist-linguist Boris Vildé, and included anthropologists Germaine Tillion and Anatole Lewitzky as well.   On 19-20 May 1942, Leiris had a nightmare of Lewitzky's execution, dreaming that his memories would be published (GHRENASSIA 1987: 239).   Leiris also dreamed that he saw Boris Vildé in a café, lifting his glass to German officers, crying out "Heil de Gaulle!" in an act of parodic resistance (LELONG 1987: 333).

10 For a discussion of this itinerary over a five-year period, see GRIAULE (1983: 174-175).   See also DIETERLEN (1971).

11 For a discussion of different sigi "dialects" in the villages of Banani, Oro Sono, Kunnu and Kabage, see LEIRIS (1992: 33).   As a combination of Mande and Voltaic elements, the sigi language involves complex considerations of origins, migrations and borrowings from Mossi, Malinke, and Bambara language areas, as well as possible Arabic influences in articles like al-, reflecting the past through linguistic forms and archaisms as well as the problematic genealogy of the Dogon themselves (ibid.: 25-44, 404.)

12 For extended discussions and studies of the textual status of Mande and Cameroonian epic genres, see AUSTEN (1995, 1999).

13 See LEIRIS (1992: 198, lines 1, 4, 5, 6, 13- 17).   I have left out the sigi texts and their initial morpheme by morpheme translation because they are not necessary in illustrating the bodily orientation of verbs.

14 For an analytical formulation of corporeal fields, see HANKS (1990: 91-95).

15 "L'emploi de ces locatifs prête aisément à amphibologie, aucune flexion n'indiquant le rôle joué dans la proposition par un nom quelconque qui se trouve placé immédiatement avant un locatif et, par ailleurs, le sujet ou complément du verbe pouvant n'être que sous-entendu, ce qui est une autre cause de confusion" (LEIRIS 1992: 422).

16 Her own, unpaginated citation is to her father (GRIAULE 1983).

17 In this respect, the Bambara and Dogon notions of nàma are just like the Yoruba idea of àse.

18 The not so subtle implication of this claim is that the work of his students and others on his team, such as Leiris, Schaeffner, Leiris, and Paulme, remain limited to this superficial level, whereas his own Masques becomes the first stage of a deepening series of further studies.

19 This logic of containment, by which a relation between elements is "counted" as an additional element is also found in the Yoruba proverb, "Two Ogboni make three" ("gbóni méji ó d'éta"), illustrating how the "third" party in the relationship is the secret (awo) that binds them.

20 See also PAULME (1940: 314-320).

21 Note the male body in GRIAULE & DIETERLEN (1954: 97-98) with female parts, and the female body in GRIAULE (1965: 94-95).

22 See TAIT (1950: 176-177) for the full list of villages.

Haut de page

Table des illustrations

Légende DIAGRAM 1. Dogon Village Scheme (from Griaule 1965: 95)
Fichier image/jpeg, 131k
Légende DIAGRAM 1. Dogon Village Scheme (from Griaule 1965: 95)Cara ïbes noirs et negros franceses : des îles au continent
Fichier image/jpeg, 222k
Haut de page

Pour citer cet article

Référence électronique

Andrew Apter, « Griaule's Legacy: Rethinking "la parole claire" in Dogon Studies », Cahiers d’études africaines [En ligne], 177 | 2005, mis en ligne le 01 janvier 2007, consulté le 29 mars 2017. URL :

Haut de page

Droits d'auteur

© Cahiers d’Études africaines

Haut de page