The System of Yoruba Oral Genres
How Yoruba verbal art is classified into named genres by vocal manner rather than subject, who performs each and on what occasion, and why the European category "oral literature" fits the material imperfectly.
Yoruba verbal art is organised into named genres, and Yoruba speakers use those names precisely. Ìjálá, ẹ̀sà, rárà, ẹkún ìyàwó, ìrèmọ̀jé, ọfọ̀, ìyẹ̀rẹ̀ Ifá, àlọ́ and orin are not a scholar's filing system imposed after the fact; they are categories the tradition itself operates with, and a performer who is asked for one and delivers another has made an identifiable mistake. The surprising part, and the thing that most distinguishes this system from European genre thinking, is what the categories are based on. They are based mainly on how the words are voiced, not on what the words are about.
Classification is by vocal expression, not by content
The same body of material can travel between genres. A passage of oríkì praising a lineage can appear in a hunter's ìjálá, in a masquerade's ẹ̀sà, and in a bride's ẹkún ìyàwó, and in each case it is genuinely that genre rather than a borrowing. Babalola established this for ìjálá and ẹ̀sà, observing that the same poem can occur in different chant modes and be called by different names . Akinyemi puts the consequence directly: the same material can occur in one or another mode of vocalisation, and it is by the mode rather than by the material that the performance is labelled .
This means content is not a reliable classifier, and neither is the identity of the performer. Ìjálá is traditionally the hunters' genre associated with Ògún, yet Yoruba Christians and Muslims incorporate ìjálá chanting into their own worship, and the genre is still recognised as ìjálá . Ògúndáre Fóyánmu, the best-known ìjálá chanter of the recorded era, is known for filling his chants with contemporary political material including a chant on the Nigerian tax war of 1969, and this does not make the performance something other than ìjálá . Neither the subject nor the singer settles the question.
What settles it is the voice. There is scholarly consensus, running from Ọlátúnjí through Ọlábíyì Yai, that the genres of Yoruba oral poetry are best distinguished and classified on the basis of vocal expression . When a Christian or Islamic musician imports a Yoruba oral genre into religious worship, listeners identify the genre not from the content and not from who is chanting, but from the distinctive vocal manner attached to it .
An acoustic study published in 2022 measured this for the first time. Akinbo, Samuel, Alaga and Akingbade had a trained chanter perform an identical text three ways, in ordinary speech, in ìjálá and in ẹ̀sà, and measured nine acoustic parameters . The findings were specific:
Ìjálá is chanted in vibrato and ẹ̀sà is not. This was already the descriptive claim in Babalola's 1963 thesis, and the measurement confirms it . The vibrato rate in ìjálá falls between 1.6 and 2 Hz, which is well below the 4 to 7 Hz typical of singing vibrato. The authors read this as consistent with vocal tremor, rather than musical vibrato, being the historical origin of the ìjálá voice . That matters because two of the four Yoruba accounts of ìjálá's origin trace its vocal manner to Ògún's aged voice and to his drinking . The acoustic profile is compatible with a tradition that says the genre's sound began as an old man's unsteady voice.
Ẹ̀sà is marked by vowel insertion and lengthening. Specifically, a triply long vowel [ooo] at the end of the first poetic line is a recurrent feature of ẹ̀sà regardless of subject matter , a point established by Olajubu's work on the genre .
Both poetic modes are pitched higher and louder than speech. Fundamental frequency was significantly raised in both, and both showed higher intensity and higher cepstral peak prominence than speech mode, indicating greater vocal effort . Ẹ̀sà was found to be higher in F0 than ìjálá across all three tones.
This raises a question the genres have to solve. Yoruba is a tone language in which pitch carries lexical meaning, so raising and shaping pitch for poetic effect risks destroying the words. Every Yoruba chanted genre is a negotiation between the melodic demands of the mode and the linguistic demand that tone contrasts survive. Chanting sits between speech and song for this reason, and the Yoruba genres are accordingly described as semi-musical verbal arts, the performance moving closer to speech or closer to music depending on genre and mood . The tonal system is treated fully in yoruba-phonology-and-tone.
The genres, who performs them, and when
A useful published classification sorts thirteen chant forms by how tightly the text is bound to its occasion and how much the performer may alter it .
Context-restricted or frozen forms. Ìyẹ̀rẹ̀ Ifá, the divination chant of the Ifá priest. Èṣù pípè, invocatory chant. Ọfọ̀, incantation. Ìrèmọ̀jé, the hunters' funeral dirge. Okú pípè, funeral dirge. In these the text is not the performer's to edit, and the efficacy is held to be tied to the formulaic structure, so that adding, subtracting or rearranging voids it .
Transitional forms, once sacred and now substantially secular. Ṣàngó pípè and Ọya pípè, the invocatory chants of those òrìṣà. Ìjálá. Ẹ̀sà egúngún .
Originally secular forms. Oríkì orílẹ̀, the lineage praise chant. Rárà. Ẹkún ìyàwó. Etíyerí, the satirical chant .
The frozen and malleable distinction is the more useful axis. In the malleable genres the performer determines the text, and the shape of what is actually delivered is decided by the audience present and by the performer's own invention, so that the performance text is a heavily altered version of anything rehearsed . In the frozen genres the words are the instrument and a fault in rendition is held to make the utterance ineffective. Ọlátúnjí describes ọfọ̀ in exactly these terms, as a restricted poetic form, cultic and mystical in its expectations, involving the uttering of words according to a formula and in a set order . That is the subject of file 07.
Occasion and performer, in outline:
| Genre | Performer | Occasion | Vocal manner |
|---|---|---|---|
| Ìjálá | Hunters, Ògún devotees | Ògún festivals, hunters' gatherings, entertainment | Chanted in vibrato |
| Ẹ̀sà / ìwì | Masquerade, Egúngún lineage members | Egúngún festivals, funerals | Chanted, vowel lengthening, disguised voice |
| Rárà | Professional chanters, often itinerant | Courts, ceremonies, in praise of patrons | Chanted, drawn out |
| Ẹkún ìyàwó | The bride, with age-mates | The nights before her departure to marriage | Sung and chanted, in tears |
| Ìrèmọ̀jé | Hunters | The funeral of a hunter, at night | As ìjálá |
| Ọfọ̀ | Practitioner, alone | Private, at need | Recited in a set order, low |
| Ìyẹ̀rẹ̀ Ifá | Babaláwo | Divination and Ifá rites | Chanted responsorially |
| Àlọ́ | Anyone, typically an elder to children | Evening, after the day's work | Narrated, with sung interludes |
| Orin | Everyone | Work, play, ritual, nursing, mourning | Sung |
Two cautions on the table. It compresses regional variation that is real: what is called ìwì in one town is ẹ̀sà in another, and the boundary between rárà and oríkì chanting is drawn differently in Ọ̀yọ́ and in Ìbàdàn. And the performer column states the traditional norm, not a rule about who is capable.
Why "oral literature" fits imperfectly
The phrase is serviceable and this corpus uses it, but four things in it are wrong for the material.
"Literature" imports the wrong object. In a written tradition the text is the artefact and the performance is a realisation of it. Here there is no artefact standing behind the performance in that way. A printed ìjálá is a transcript of one chanter's rendition on one night, not the work of which that night was a performance.
"Oral" suggests the absence of writing rather than the presence of voice. It names the tradition by what it lacks. But the vocal manner is the classifying feature of the whole system, as above, and calling it oral captures none of that. Ọlábíyì Yai has argued against exactly this framing, and against the transposition of European literary categories onto Yoruba verbal art .
The Western genre map does not align. There is no Yoruba category corresponding to "poetry" that would group ìjálá with ẹsẹ Ifá and exclude prose narrative. The tradition's own cuts fall elsewhere, along lines of vocalisation and occasion.
It obscures function. Ọfọ̀ is not intended to be appreciated. Oríkì is addressed to a subject who is present and expected to respond. Ẹfẹ̀ songs name living people in order to correct their behaviour. To call this body of practice literature is to file a set of social actions under aesthetics.
Against that, the phrase is not simply wrong. This material is verbal art, composed with formal ambition, judged by demanding standards, and preserved because people valued it as made language and not only as social machinery. Babalola's account of how ìjálá performances are judged is an account of criticism . The reasonable position is that "oral literature" names a real object badly, and that the Yoruba genre names are the accurate vocabulary.
Barber's argument: text and performance are not opposites
Karin Barber's central theoretical contribution, developed from her fieldwork in Òkukù, is the strongest available answer to the question of what an oral text is. Her argument runs against two positions at once.
Against the older literary approach that treated an oral performance as a degraded version of a text, she holds that the performance is where the work happens. Against the performance-theory reaction of the 1960s onward, which fought a successful battle against a scriptocentric view of orality and in doing so came close to a witch-hunt against the very concept of "text", she holds that oral performance nonetheless constitutes something text-like, and must .
Her reasoning is that oral traditions cannot be pure evanescence, or they would not be traditions:
"There is a performance, but it is a performance of something. Something identifiable is understood to have pre-existed the moment of utterance."
Traditions are shareable and handed down because they are constituted precisely so as to be detachable from the immediate context . The technical term she takes up from linguistic anthropology is entextualization, the process of rendering an instance of discourse as text, detachable from its local context .
The interesting move is what follows. Entextualization is usually associated with third-person, remote-past, myth-like discourse: language stripped of pointers to the here and now. But African praise poetry is the opposite of that. It is vocative and second-person, it addresses a subject who is physically present, and it brings the dead into the middle of the living community . So Barber asks what entextualizing strategies a genre like this uses, and answers that they work by rendering discourse object-like: by making it the focus of exegesis, and by presenting it as quotable, so that the words are heard as having pre-existed this utterance and as continuing after it .
Two mechanisms carry this.
Obscurity that demands exegesis. The characteristic device of Yoruba praise poetry is the nominalised epithet, a whole sentence compacted into a name. Barber's example is from Òkukù:
Dínà-má-yà A-dóminú-kojo
- Dínà block-road má not yà budge
- A- one-who dóminú fills-the-inside-with-fear kojo of-the-coward
Block-road-not-budge. One who fills the coward with apprehension.
Translator: Karin Barber
Dínà-má-yà is a nominalisation of the sentence Ó dínà, kò sì yà, "he blocked the road and he didn't budge" The conversion strips the pronoun and the verbal aspect, so an event performed by an agent at a time becomes a timeless attribute. English has no comparable operation and must either hyphenate, as Barber does, or unpack into a clause and lose the compactness. The corpus file yoruba-etymology-and-word-formation describes this nominalising process as a general feature of the language.
These two epithets belong to Wínyọmí, a nineteenth-century hunter of Òkukù, and to his descendants . From the words alone you cannot recover why. The narrative is held in a separate genre, the ìtàn: Wínyọmí sheltered a man who had killed someone during a masquerade outing; the dead man's relatives marched on Òkukù to demand him; Wínyọmí came out, stood in the road and said he would close his eyes and when he opened them they would all have gone, and the crowd fled .
The epithet, then, is deliberately incomplete, and the completion lives outside it. Barber's observation about who holds which half is important: in Òkukù, oríkì are performed mostly by women and ìtàn are recounted mostly by elderly men, so the production of textual meaning is divided between two institutionally separate genres and two groups of people . Neither half is the whole text.
Quotation. The second mechanism is explicit marking of words as pre-existing. Barber's example, from a praise performance:
"Ó gbọ́ Ifá ó yalé Ó gbọ́hìnkùlé mọ̀ye odù tó hù 'Àìgbọ́fá là á wòkè, Ifá kan ò sí párá'" Ló tọ́ Babaa Fárónkẹ.
- Ó gbọ́ Ifá he hears Ifa ó yalé he turns-into-the-house
- Ó gbọ́hìnkùlé he from-the-backyard mọ̀ye odù knows-the-figure-of-odu tó hù that has-emerged
- Àìgbọ́fá not-knowing-Ifa là á wòkè is-what-we look-up Ifá kan ò sí párá no Ifa is on-the-rafters
- Ló tọ́ is-what-is-worthy Babaa Fárónkẹ Father-of-Faronke
"He hears the chink of the divining chain, he stops to come in, from out in the yard he already knows what figure has emerged. 'Not knowing Ifá, we gaze up, but there's no Ifá in the rafters.'" This is what Father of Fárónkẹ is worthy to be called.
Translator: Karin Barber
The first two lines sound like statements about a man. The third is a proverb, and it is used with what Barber calls characteristic Yoruba inversion: the proverb describes people who do not know Ifá and look uselessly upward, and by quoting it the performer implies that her subject is the opposite . Then the fourth line reaches back and rounds up everything just said, declaring it a thing the subject is worthy to be called. The whole passage is retroactively converted from description into attribution, and simultaneously marked as quoted, as belonging to a repertoire that can be applied to any diviner who merits it . English can carry the sense but not the grammatical gesture of ló tọ́, which does the rounding-up in three syllables. This is also the clearest single illustration of how ọwe behave inside performance, treated in file 10.
Barber's conclusion is that these are not opposed processes:
"Entextualization is not the opposite of emergent performance, but rather its alter ego; they proceed hand in glove with each other and are the condition of each other's possibility. For text must be treated as the object of attention, by exegesis and by being quoted in new contexts of utterance, in order to attain meaning; while a performance that was truly ephemeral would be a performance of nothing."
The consolidation of compact, quotable, partly obscure chunks is precisely what makes fluid performance possible. Because each fragment is self-contained and points outward to its own narrative hinterland, fragments become mobile with respect to each other, and a performer can select, reorder, expand and combine them freely .
The audience is inside the performance
This follows directly, and it is where the Yoruba genres depart most sharply from a literary model.
An oríkì performance is a stringing together of autonomous fragments which could in principle be delivered in any order or combination . What holds it together is not plot or argument. It is the presence of the addressed subject, in whom all the attributions converge, plus the performer's skill at throwing out temporary links between fragments based on similarities of sound or sense .
The performer engages the addressee in an intense dyadic relationship with unwavering eye contact, exhorting, blessing and thanking, and will switch to a new subject when a more important person enters the performance space . The addressee is not a spectator. They are the organising principle, and they are expected to respond, materially included: praise performance has always had an economic dimension, and the performer is rewarded by the praised.
There is a further consequence Barber draws for oríkì specifically. Because this is a culture of competition between self-aggrandising big men, and because the more oríkì heaped on an addressee the greater their aura relative to rivals, performers do not confine themselves to an authorised corpus for each subject. They raid other subjects' oríkì, and they raid other genres, proverbs, riddles and Ifá verses, for material to add to the flow . The result is a pervasive intertextuality in which borrowed elements remain partly unassimilated, casting what she calls a haze of quotedness over the whole field of Yoruba verbal art .
That haze is the reason this section is organised by genre but keeps pointing sideways. The genres are real and distinct in their vocal manner and their occasion, and they are continuously trading material.
What this means for reading the texts in this section
Three practical consequences.
A printed text is one performance. Not the poem. Where a source names the chanter, the place and the date, that is a considerable part of what makes it evidence, which is why this corpus records those details when they exist and says so when they do not.
Incompleteness is often the text working correctly. A line that alludes in four words to a story you do not know is not corrupt. The expansion was always held elsewhere, sometimes by different people in a different genre, and where the ìtàn has been lost the line survives as a genuine fragment. Scholarly interpretation of such lines is reconstruction and should be labelled as such.
Transcription loses the classifier. If genres are distinguished mainly by vocal expression, then the printed page removes the feature that makes a text the genre it is. The words of an ìjálá and the words of an ẹ̀sà can be identical; the vibrato and the lengthened [ooo] are what differ, and neither survives transcription. This is the strongest argument for the recordings listed in file 14 and in sources-audio-visual, and it is why a reader who has only read this material has an incomplete object.