African Mythology and Oral Tradition: Reading the Form

There is a question that comes near the end of almost every reading. A hand goes up somewhere in the middle rows, and someone asks the novelist where the magic came from — as though African mythology were a quarry, and the book a thing chipped out of it. The novelists I trust do not answer with a source. They answer with a name. A grandmother in Kumasi. A jeli in Kita who sang the same king's story twice in one week and did not sing it the same way twice. An uncle who told the tortoise story four different ways depending on which child in the room needed correcting.

If you grew up anywhere near this, you already know the difference between a myth and a telling. You have watched a story change mid-sentence because someone walked into the room, or because the teller wanted something from you and adjusted the ending to get it. What you may never have been handed — because nobody sat you down and gave it a name — is the vocabulary for the thing you already noticed. These traditions have a form. The form is doing as much work as the content, sometimes more.

That is the argument here. Most writing about the continent's myth gets the nouns roughly right and the shape entirely wrong, and the shape is where the meaning lives.

Is African mythology written down anywhere?

Yes — extensively, and in more than one sense. Hundreds of oral performances have been recorded in the field, transcribed in the original language, translated, and published as books you can borrow: the Mande Sunjata epic, the Nyanga Mwindo epic, thousands of Yoruba Ifá verses, Zulu praise poems, Akan tales collected in the 1920s. Those books are not the tradition. They are photographs of particular nights, with a named performer standing in the frame.

The question also carries an old assumption worth setting down gently: that the continent had speech and Europe had writing. It is not true. Ge'ez has been written in Ethiopia and Eritrea for well over a millennium. Ajami — Arabic script adapted to write Hausa, Wolof, Fulfulde, Swahili — carried poetry, law, and correspondence across West and East Africa for centuries. Timbuktu's family libraries hold tens of thousands of manuscripts. Nsibidi, a system of graphic signs used among the Ejagham, Efik and Igbo of southeastern Nigeria and Cameroon, encoded meaning without borrowing anyone's alphabet. The Vai syllabary of Liberia was devised in the 1830s by Vai men who wanted their own script and made one.

So the honest version is this: writing and orality have lived side by side here for a very long time, and oral performance was never the poor relation waiting for paper.

What most people do

What most people do — writers, screenwriters, game designers, and a fair number of us reading — is treat the continent's myth as a props cupboard. You go in, you take a deity, you file the serial number off, you give the character a power and a temper, and you close the door. Ọ̀ṣun becomes a water power-up. Èṣù becomes a villain, which is a missionary-era mistranslation with a long tail — he is the Yoruba òrìṣà of crossroads, messages, and consequence, the one who carries offerings and language between worlds, and reading him as Satan says more about the translator than about the tradition.

The second habit is the imported frame. The word pantheon comes with Greek furniture attached: a fixed roster, a family tree, a mountain, a canon everyone agrees on. Yoruba tradition counts òrìṣà differently in different towns. Akan cosmology does not map onto Olympus and does not want to. When a book announces "the African pantheon" and seats Yoruba òrìṣà beside Egyptian neteru and a Zulu praise-name, it has not built a world; it has built a shelf, and the shelf is the giveaway.

Third is the flattening. A continent of more than a thousand languages gets one belief system, one drum sound, one red-dust establishing shot. The Swahili coast and the Sahel and the Congo basin are not variations on a theme. They are different intellectual traditions with different questions. The Akan think hard about the person as a composite of inherited elements. Ifá is a divination system with a memorized literary corpus that runs to hundreds of thousands of lines. These are not the same conversation.

Fourth — and this one is subtler, because well-meaning books do it — is the aside. A novel is moving, and then it stops so a character can explain to another character something both of them have known since childhood. As you know, the ancestors are believed to intercede... That sentence is not written for the person in the scene. It is written for a reader assumed to be standing outside. If you have ever read a novel set in Lagos or Kingston or Peckham and felt the prose turn its head away from you to address someone else in the room, you have felt this. Fiction that reflects a worldview does not narrate it. It assumes it and gets on with the plot.

And fifth: ancient. Applied to a shrine that opened last year, to a diviner who has a phone, to a rite performed last Tuesday. Ancestor veneration is not a survival. Ifá divination is practised now, by people with jobs. The past tense is a choice, and it does a quiet kind of harm.

None of this is malice. Mostly it is a reader who had only listicles and wiki summaries to work from, writing the book they were trained to write.

What the evidence suggests

Go to the recorded performances — the field transcripts, the ones with a named teller and a date — and the first thing that becomes obvious is that oral tradition is not a text that happens to be spoken. It is performance, and performance has rules that the printed summary throws away.

Variation is the design, not the damage. Among the Mande peoples of Mali, Guinea, Senegal and Gambia, the story of Sunjata Keita — the thirteenth-century founder of the Mali empire — belongs to the jeliw, hereditary bards whom French colonial writing called griots. D. T. Niane's widely read 1960 version came from Djeli Mamoudou Kouyaté of Guinea. The version John William Johnson recorded from Fa-Digi Sisòkò in 1968 runs differently: different emphases, different episodes, a different weighting of lineages. Neither is corrupt. A jeli performs for a particular audience, in a particular town, where particular families are listening for their own names. The story is stable in its bones and mobile in its flesh, and the mobility is a skill, not a slippage.

Audiences are inside the form. Among the Akan of Ghana and Côte d'Ivoire, an evening of Anansesɛm — Ananse stories, after the spider — is not a monologue. Listeners interrupt with songs, called mmoguo, that they choose themselves; the story pauses and absorbs them. The Nyanga epic of Mwindo, from eastern DR Congo, was recorded in the 1950s from the narrator Shé-Kárisi Rureke across a stretch of days, with singers answering him and a drum keeping the room awake. A novel that renders any of this as one voice speaking into silence has translated a conversation into a lecture.

The traditions make their own genre distinctions, and they are sharper than ours. Akan tellers have long opened a tale with a formula that translates roughly as we do not really mean, we do not really mean — a disclaimer, stated up front, that what follows is not being offered as fact. That is a culture drawing a clean line between a made thing and a historical claim, and it makes nonsense of the idea that people who tell stories about a spider cannot tell stories from history. The same community holds genealogy, drum history and stool history to entirely different standards of accuracy.

Some of it is a literature held in memory. Ifá, the Yoruba divination system practised in Nigeria, Benin and across the Atlantic, is organized into 256 odù, each with a body of verses attached. A babaláwo — an Ifá priest, the title means roughly "father of secrets" — spends years committing them to memory. Wande Abímbọ́lá's published expositions from the 1970s put the scale of that corpus in front of a general reader for the first time; UNESCO later recognised Ifá as intangible cultural heritage. Whatever else it is, it is one of the larger bodies of poetry in the world, and it is not on a shelf.

Praise is a form in its own right. Yoruba oríkì are strings of attributions — names, epithets, fragments of history — attached to a person, a lineage, a town, or an òrìṣà, and performed most often by women. Karin Barber's study of oríkì in a Yoruba town describes them as deliberately non-narrative: they accumulate rather than resolve. Zulu and Xhosa izibongo work by comparable logic, delivered by an imbongi, with new lines added as the subject's life goes on. If you are looking for the aesthetic that most distinguishes this material from the European epic, it is here — meaning built by accretion rather than by plot.

Versions differ on nearly everything in that table, which is the point. A tradition that changes with the room is not less reliable than one that does not. It is answerable to a live audience, which is a harder test than print.

What I actually do

I am a reader and a researcher. I am not an initiate of anything, and I write toward these traditions rather than from inside them. That position sets most of my rules.

I go to the performance, not the summary. If a story matters to a piece I am writing, I find a transcript with a named teller on it — Rureke, Sisòkò, Kouyaté, the people Rattray recorded in Asante in the 1920s — and I read the whole thing at its own pace, digressions included. Summaries strip out exactly what I need: repetition, the pause for song, the aside to a listener, the moment the teller names someone's grandfather. The plot survives summarising. The form does not.

I name the version. In my own prose I write in the telling Rureke gave, or among the Yoruba of Ọ̀yọ́, rather than the myth says. It costs four words and it removes the fiction of a continental canon. It also protects me: if I have only one version, saying so is more honest than implying I have surveyed them all.

I steal form before content. This is the useful move and almost nobody makes it. If I want a novel to feel like Anansesɛm, the thing to carry over is not the spider — it is the opening disclaimer, the audience that sings back, the ending that hands the story on to someone else. Structure travels better than character. It also cannot be accused of theft in the way lifting a sacred figure sometimes can.

I cut the aside. If two characters both know what an ancestor is, neither of them explains it. Meaning gets carried by consequence: someone breaks a rule and the room goes quiet, and the reader who knows why feels seen, while the reader who does not learns from the silence. Anything genuinely obscure goes in a note at the back, where a reader can choose it. A book that explains itself in the margins is a book addressed to a stranger, and there are enough of those.

I leave the closed things closed. A diviner can appear in a scene, and I can write what the consultation costs the person who asks for it, what it means to the household, what happens after. I do not transcribe the verse. I do not write the offering, the invocation, or the sequence of a rite, and I do not write anything a reader could follow as instructions. This is not squeamishness. Ceremony belongs to lineages and to the people trained in them, and a novel that hands it out as a recipe has taken something it was not given. The meaning is mine to write about; the manual is not mine to write.

I keep living practice in the present tense. Ifá is consulted. Ancestors are honoured. Vodou is practised, in Haiti and Benin and Brooklyn, by people who will read your book. The past tense turns a neighbour into an exhibit.

I ask what the story was for. Ananse is not a hero and not a villain; the tales run him through both and often let him lose in a way that is funny rather than just. That ambivalence is the moral machinery, and flattening him into a lovable rogue or a cautionary example breaks it. Before I use a story, I try to work out what argument it was making to the people who kept it, and whether my book is making that argument or quietly reversing it.

Names and terms

• Jeli (pl. jeliw, Maninka; often "griot") — a hereditary Mande specialist in history, praise, genealogy and music.
• Anansesɛm (ah-nan-seh-SEM) — Akan tales centred on Ananse the spider, told in the evening with songs from the listeners.
• Oríkì (oh-REE-kee) — Yoruba praise poetry: strings of names and attributes belonging to a person, lineage, town, or òrìṣà.
• Odù — the 256 divisions of the Ifá corpus, each holding a body of memorized verse.
• Babaláwo — an Ifá priest and diviner; the word means roughly "father of secrets".
• Izibongo — Zulu and Xhosa praise poems, delivered by an imbongi, the praise poet.
• Ajami — Arabic script adapted to write African languages such as Hausa, Wolof, Swahili and Fulfulde.
• Nsibidi — a system of graphic signs used among the Ejagham, Efik and Igbo of southeastern Nigeria and Cameroon.

What this piece didn't answer, and where to look next

It did not answer the hardest question, which is who holds the right to tell what. I have given you a working position — describe, don't instruct; name the teller; leave the closed things closed — and a working position is not a settled ethics. Nor did I touch what a novelist owes a living performer whose recorded telling made their book possible, which is a question about money as much as credit, and one the publishing industry has mostly declined to hear. I have said almost nothing about the Amazigh traditions of North Africa, the Ethiopian qene tradition with its deliberate double meanings, the Somali poetic culture that argues politics in verse, or the Swahili coast, where Islam and older practice have been in conversation for a thousand years. And I have skipped the case where a tradition's own custodians disagree with one another, which happens constantly and which no outside method resolves.

Where to look next: Ruth Finnegan's Oral Literature in Africa, the standard survey, is freely readable online. Isidore Okpewho's African Oral Literature gives you the performance-centred approach in one volume. Then go to primary recordings rather than retellings — Biebuyck and Mateene's The Mwindo Epic, Johnson's The Epic of Son-Jara, Niane's Sundiata, Karin Barber's I Could Speak Until Tomorrow on oríkì, Abímbọ́lá on the Ifá corpus. Read the introductions; that is where the fieldworkers admit what they changed.

And then look closer to home. The last serious archive is the one with your surname on it: the relative who still knows the family's praise lines, the phrase your grandmother used that nobody translated. Ask while there is someone to ask.

Akan tellers have a formula for handing a story on at the end of the night: this is my story; if it be sweet, or if it be not sweet, take some elsewhere, and let some come back to me.