How Yoruba History Is Actually Known
The kinds of evidence historians use to reconstruct the Yoruba past, what each kind can and cannot prove, and how one 1921 book shaped everything written since.
Yoruba history before the nineteenth century is reconstructed from four separate bodies of evidence that rarely agree cleanly: oral tradition performed and transmitted inside Yoruba institutions, archaeology and the radiocarbon dates that come with it, historical linguistics, and written sources produced by Arabic-literate and European outsiders. Each answers a different kind of question. Oral tradition tells you what a community says about who it is and where its authority comes from, which is a historical fact about the community even when the events described cannot be verified. Archaeology gives you dated material remains but is nearly mute about names, dynasties, and motives. The honest position is that these bodies of evidence are combined by argument, not by addition, and that competent scholars reading the same evidence reach different conclusions.
This file explains the machinery. It exists because a reader who does not understand how a claim about the Yoruba past was produced cannot tell the difference between a well-dated archaeological finding, a plausible reconstruction, a contested reading of a tradition, and a colonial-era assertion that has been repeated so often it now sounds like fact. Those four things appear side by side in most popular accounts, undifferentiated.
Itan: oral tradition as a Yoruba genre, not as raw data
The Yoruba word for a historical account is ìtàn, from the verb tàn, to spread out or unfold, so an ìtàn is literally something spread out for viewing . Ìtàn is not a single genre. It covers dynastic narrative, town-founding accounts, the narrative portions of ẹsẹ Ifá, and the aetiological stories attached to òrìṣà. Alongside it sits oríkì, praise poetry attached to persons, lineages, towns, and deities, which Karin Barber has shown to be the deepest single repository of historical reference in Yoruba verbal art precisely because it is not narrative . Oríkì accumulates in fragments, each attached to a name, and it preserves epithets and allusions long after the narrative that explained them has been lost. Barber's argument is that oríkì works by juxtaposition rather than sequence, so it retains material that a story would have smoothed away for the sake of coherence .
Three layers:
Ọ̀yọ́ ni ilé, ìlú mìíràn ni ọjà.
Ọ̀yọ́ / it-is / house, town / other / it-is / market.
Ọ̀yọ́ is home; every other town is a marketplace.
Translator: rendered here from a widely circulated Ọ̀yọ́ oríkì formula; the rendering is my own from the transcribed Yoruba and should be treated as medium confidence, not as a published scholarly translation.
The line is not a description, it is a claim. It sets ilé, the permanent dwelling and by extension the ancestral seat, against ọjà, the market, which in Yoruba idiom carries the sense of a place one visits and leaves. The proverb Ayé lọjà, ọ̀run nilé, the world is a market and heaven is home, uses the same pairing, so the Ọ̀yọ́ formula is borrowing an existing cosmological contrast to make a political claim about primacy. English can carry the words but not the fact that the audience already knows the other proverb. This is exactly the kind of material that is historically valuable as evidence of Ọ̀yọ́ self-assertion and worthless as evidence about settlement chronology.
The critical point for method: ìtàn and oríkì are performed inside institutions that have stakes. The arọ̀kin, the corps of royal bards at the Ọ̀yọ́ court, were court functionaries whose recitations legitimated the reigning line . Traditions of origin are, in Robin Law's formulation, political charters, and their content shifts with the political needs of the moment . Law's study of Odùduwà traditions argues that claims of descent from Odùduwà spread outward over time as a device by which ruling houses asserted a rank they did not previously claim, which means the distribution of the tradition tracks the diffusion of a political idea rather than the movement of a population .
None of this makes oral tradition useless. It makes it evidence of a specific kind. It is strong evidence for how a polity understood and justified itself at the time of recording, moderate evidence for sequences of rulers within a few centuries of the recording, and weak evidence for absolute dates or for events many centuries earlier.
Samuel Johnson and the shape of everything downstream
Samuel Johnson (1846 to 1901) was born in Freetown, Sierra Leone, to Yoruba parents who had been liberated from a slave ship, and he was a great-grandson of Aláàfin Abíọ́dún of Ọ̀yọ́ . He returned to Yorubaland as a child, trained as a Church Missionary Society agent, was ordained deacon in 1880 and priest in 1888, and was stationed at Ọ̀yọ́ from 1881 . He was an active participant in the peace negotiations that ended the nineteenth-century wars, not a detached observer of them. He completed his manuscript in 1897 and sent it to the CMS in London, where it was lost. His brother, the physician Obadiah Johnson, reconstructed the work from Samuel's notes; the reconstructed manuscript was aboard the ship Appam when it was captured by a German raider during the First World War and ended up in America, and wartime paper shortages delayed printing further. The History of the Yorubas from the Earliest Times to the Beginning of the British Protectorate finally appeared in 1921, twenty years after the author's death .
Johnson's stated motive was that "educated natives of Yoruba are well acquainted with the history of England and with that of Rome and Greece, but of the history of their own country they know nothing whatever" The book became, and remains, the most cited single volume on the Yoruba . That is the problem as much as the achievement.
Robin Law's 1984 article in History in Africa is the standard critical treatment . Law attacks the assumption that what Johnson wrote down is unproblematically identical with what he heard . Three distinct issues sit inside that: Johnson used written as well as oral sources and did not always distinguish them; the surviving text is a reconstruction by a second hand from notes, so we cannot always tell Samuel's words from Obadiah's; and the oral traditions Johnson collected had themselves already been reshaped by Islamic and Christian influence and by nineteenth-century political circumstance before he reached them . Law shows, for instance, that Yoruba origin traditions of the period carry material traceable to Islamic legend, including the Nimrod and Abraham cycle, which means the tradition Johnson recorded as ancient was in part a nineteenth-century composite .
The specific distortion that matters most is perspectival. Johnson was an Ọ̀yọ́ man, an Anglican clergyman, and a descendant of an Aláàfin, writing in and for Ọ̀yọ́, and his book treats Ọ̀yọ́ as the centre and other Yoruba polities largely as they relate to Ọ̀yọ́. Its narrative frame, in which Ilé-Ifẹ̀ is the spiritual cradle and Ọ̀yọ́ the political apex and the other kingdoms are branches, became the default structure for Yoruba history for most of the twentieth century. It has been corrected in three main ways.
First, historians of the eastern and coastal kingdoms, including Ijẹ̀bú, Ìjẹ̀ṣà, Èkìtì, Ondó and Ọ̀wọ̀, have shown these polities had institutional histories not derivable from Ọ̀yọ́ and in several cases older than the Ọ̀yọ́ imperial period. Second, archaeology has established a dated chronology for Ilé-Ifẹ̀ that is independent of dynastic tradition and that places Ifẹ̀'s florescence centuries before Ọ̀yọ́'s imperial expansion. Third, J. D. Y. Peel and others have shown that "Yoruba" as a single ethnic label covering all these polities is itself a nineteenth-century product, so a book that narrates a unified Yoruba people from the earliest times is projecting a recent category backwards .
Read Johnson, but read him as a primary source: a document produced by a specific man in a specific position at a specific moment, invaluable for late eighteenth and nineteenth century Ọ̀yọ́ where he had access to court informants and lived through the events, and to be handled with care everywhere else.
Archaeology: what dated material can and cannot say
Archaeology in Yorubaland gives absolute dates through radiocarbon and thermoluminescence, physical evidence of production technologies, and evidence of settlement scale and layout. It does not give names, and it establishes chronology for sites, not for dynasties.
The core body of work at Ilé-Ifẹ̀ begins with Frank Willett, who conducted rescue excavations at Ita Yemòó in December 1957 and January 1958 after labourers preparing land for construction uncovered copper-alloy castings, and who ran further campaigns through 1963 . Willett's Ita Yemòó dates fall in the eleventh to thirteenth centuries CE; his work at Òrun Ọba Adó produced dates from the eighth to twelfth centuries . A 2021 study reconstructing Willett's excavation records notes that the site's fame contrasts sharply with how little of its archaeology was ever published in detail, and that unpublished Willett records were needed to recover the stratigraphy . That is a general condition of Ifẹ̀ archaeology: several of the most famous finds were recovered in rescue conditions or by chance, with weak stratigraphic control, which limits how tightly they can be dated.
The most secure recent work is on glass. James Lankton, Akin Ige and Thilo Rehren proposed in 2006 that the high-lime, high-alumina glass found at Ilé-Ifẹ̀ was made locally rather than imported and remelted, which had been the prior assumption . Abidemi Babalola's excavations at Ìgbó Olókun between 2010 and 2012 recovered over 12,000 glass beads together with hundreds of glass-encrusted crucible fragments and other production debris . The compositional argument is decisive in a way that stylistic arguments are not: this glass runs roughly 10 to 20 percent alumina and 12 to 20 percent lime, a signature not found in European, Middle Eastern or Asian glass, and it maps onto locally available high-alumina pegmatite sands combined with a lime source in snail shell . Primary glass production in sub-Saharan Africa in the early second millennium CE is therefore established on chemistry, not inference.
What archaeology cannot do is attach these findings to the persons named in tradition. No excavated object at Ifẹ̀ carries an inscription naming a ruler. Attributions of particular castings to particular Ọ̀ọ̀ni are art-historical arguments from iconography and style, not archaeological findings, and they should be labelled as such.
Linguistics
Yoruba belongs to the Yoruboid cluster, whose other members are Ìgálà and Itsẹkiri, within the larger Volta-Niger grouping of Niger-Congo . Earlier classifications placed it in Kwa and later in Benue-Congo; the current mainstream placement is Volta-Niger alongside Igbo, Edo and the Gbe languages . The reconstructed ancestor, Proto-Yoruboid, is generally located near the Niger-Benue confluence .
Linguistic evidence supports two things well: relatedness, and rough relative sequence of divergence. It supports absolute dating poorly. Glottochronological dates circulating for the Yoruboid split, commonly given as around three thousand years before present, rest on assumptions about rates of lexical replacement that historical linguists themselves treat as unreliable for absolute chronology . Treat any specific century figure for a proto-language as an order-of-magnitude estimate. What the linguistic evidence does establish firmly is that the Yoruboid languages diverged from a common ancestor located to the northeast of the present Yoruba heartland, which is consistent with, though not proof of, Ogundiran's reconstruction of Yoruba origins in the Niger-Benue confluence region .
Dialect geography is a second and more secure linguistic tool. The internal diversity of Yoruba dialects, and the relative depth of the differences between the northwestern, central and southeastern groups, constrains reconstructions of settlement sequence. Areas of high dialect diversity generally indicate longer settlement than areas of low diversity.
Arabic and European written sources
Arabic sources for the West African interior, principally the North African and Andalusian geographical tradition, are rich for the Sahelian states and thin for the forest zone. They do not give usable early references to Ifẹ̀ or Ọ̀yọ́. Claims that particular medieval Arabic geographers described Yoruba polities should be checked against the actual text before being repeated; several such claims in popular writing do not survive checking.
European written sources begin with the Portuguese on the coast from the late fifteenth century. They are contemporaneous, which is their strength, and they are coastal, secondhand about the interior, and shaped by commercial interest, which is their limit. Dutch, English and French accounts accumulate through the seventeenth and eighteenth centuries. Hugh Clapperton and Richard Lander reached Ọ̀yọ́-Ile in 1826, producing the first European eyewitness description of the Ọ̀yọ́ capital, and their accounts are a fixed point against which oral traditions of the empire's late period can be checked.
The nineteenth century transforms the written record. Church Missionary Society agents, most of them Yoruba rather than European, kept journals which they submitted to the CMS, and these constitute an enormous, dated, near-contemporaneous corpus. Peel's Religious Encounter and the Making of the Yoruba is built on this archive and makes the case that these African agents were not merely informants but authors with substantial autonomy, and that as the first literate generation of Yoruba they were central to shaping modern Yoruba consciousness . The CMS journals are the single most important documentary source for nineteenth-century Yoruba social history. They are also mission documents, written by converts, for a missionary society, about people the writers were trying to convert, and they must be read with that frame visible.
What follows from all this
Four practical rules govern how this section is written.
Registers are kept separate. When a claim comes from tradition it is labelled as tradition. When it comes from a dated excavation the date and the method are given. When it comes from a colonial or missionary source it is identified as such and the source's position is stated.
Dates carry confidence. Archaeological dates are given with their ranges. Dynastic chronologies before roughly 1700 are reconstructions from reign-length estimates and king lists, and they are marked as such.
Disagreement is named, not smoothed. Where Law and Ogundiran, or Willett and Blier, read the same evidence differently, both positions and the evidence for each are given.
Tradition is neither promoted to fact nor demoted to legend. That an account cannot be verified as event does not make it untrue as a statement of how a people understood their own foundation, and that understanding is itself a historical object worth precise description.