Culture · Essay
The griot: a family's records kept in one person's memory
For seven centuries West African griots have carried the genealogies, marriages and quarrels of the families they serve, learned by heart and passed father to son. On what a living archive can do that a written one cannot, and what happens when it dies.Written to last.
By Confinity Editorial · 2026-08-21 · 6-minute readQuiet tools, kept out of the way.
If you belong to certain families in Mali, Senegal or the Gambia, there is a person who knows more about your ancestors than you do.
They are not a relative. They belong to a different family, a griot family, which has been attached to yours for as long as either can account for. They know the names in your line, who married whom, which branch left and why, what your great-grandfather was owed and by whom, and the praise-name that belongs to your surname and no other. They know it because their father knew it and taught it to them, and because their father's father knew it before that.
This is a professional archive with no shelves.
Griot is the French word. In Manding the word is jeli, in Wolof gewel, in Fula gawlo, and the roles they name are not identical, but the core is shared.
A griot is a genealogist, a historian, a musician and a public speaker, and the roles are one job rather than four. 
Genealogy is delivered as performance: the names are sung or declaimed, often to the kora, the ngoni or the balafon, and the music is what makes the list holdable in a mind for a lifetime.
They are also called at the points where a family's record matters. A naming ceremony, a marriage, a funeral, a dispute over inheritance or land. The griot says who these people are, going back, and their saying it is what settles the question.
And they mediate. A griot can say to a chief what nobody else can say to a chief, because a griot is outside the ordinary hierarchy of respect. That licence is part of the role and always has been.
The griot tradition is bound up with the Mali Empire, which rose in the thirteenth century and at its height was among the largest and wealthiest states in the world.
Its founding story, the epic of Sundiata Keita, has come down almost entirely through griots. It was recited for seven hundred years before it was written, and the written versions we have are transcriptions of particular griots' performances, made in the twentieth century. The best known of them was recited by Mamadou Kouyaté, of a griot family that has served the Keita line since Sundiata's own griot, Balla Fasséké.
Read that again: the family that keeps the record has been keeping the record for the same family for seven centuries. Whatever else oral tradition is, it is not casual.
The empire's capital moved, Timbuktu filled with written manuscripts, the state fell, colonial borders cut across the Mande world, and the griot lines held.
It is easy to treat oral record-keeping as a stage a society passes through on the way to writing. The griot makes a better argument than that.
A document is inert. It says the same thing to everyone, in the same words, and if you cannot read it, or cannot find it, or do not know it exists, it says nothing at all. A griot is present, addresses this family at this moment, and can answer a question. Which cousins are we to these people? Why do we not speak to that branch? A written genealogy holds the data; the griot holds the data and the reason.
A griot also knows what the family is for. The praise-names attached to a surname are not decoration. They are a compressed argument about who these people have been, deployed at the moment a person most needs to hear it.
There is a cost to that, and glossing it would be dishonest. A griot serves a family, and a record held by somebody who is paid to praise you will lean. Historians who work with griot testimony treat it as a source with a known bias, not a chronicle. The tradition is candid about this too: the flattery is part of the form.
The failure mode is obvious and it is happening.
A living archive dies. If a griot's children take other work, and many now do, seven centuries of a particular family's record ends with one funeral. There is no backup, because the person was the backup.
Recording projects exist. Griot performances have been collected, transcribed and published since the mid-twentieth century, and archives in Bamako, Dakar and abroad hold a great deal. What they hold is performances, though, not the standing capability. A recording answers the questions somebody thought to ask in 1963. It cannot answer the question your family will have in 2040.
Urban migration, formal schooling and the collapse of the patronage that paid griots have all pulled in the same direction. The tradition is not gone and in some places it is flourishing, particularly in music. The genealogical function is thinner.
Almost every society had this role and most of them lost it without noticing.
The Icelandic saga-tellers, the Irish seanchaí, the Serbian guslars, the Māori keepers of whakapapa, the Balkan singers whose technique Milman Parry studied to understand Homer. Each held a community's descent and history in memory, professionally, and each was displaced by writing, then by print, then by the assumption that somebody else is keeping the record.
Nobody is keeping it. That is the actual position of most families reading this. There is no griot, there is no genealogy book, there is no spirit tablet with the names on it, and there is no aunt who could recite four generations if asked. There is a phone with eleven thousand photographs on it and nobody who could say who is in most of them.
Not the memorisation. The griot's real innovation is that remembering was somebody's actual job, with a name, and a successor, and an occasion on which the record had to be produced in public.
Confinity is an attempt at the same arrangement without requiring a person to give their life to it: the record kept in one place, added to at the moments that matter, and handed on deliberately rather than by accident.
The griots worked out seven hundred years ago that a family's memory does not survive on goodwill. It survives because somebody is responsible for it.
The job