Ellis 1894 as a Source for Yorùbá Folktale

Ten tales in this section come from one book: A. B. Ellis, The Yoruba-Speaking Peoples of the Slave Coast of West Africa, published in London in 1894. It is the earliest substantial published collection of Yorùbá folktale narrative that this compiler was able to obtain in full, it is in the public domain, and it prints its tales at narrative length rather than in summary. It is also a book written by a serving British colonial military officer whose closing chapter ranks West African peoples in stages of evolutionary progress. Both things are true and this file sets out how the corpus handles that.

What the book is

Alfred Burdon Ellis was a Lieutenant-Colonel of the First Battalion West India Regiment. The Yoruba-Speaking Peoples is the third in a series of three, following volumes on the Tshi-speaking and Ewe-speaking peoples of the Gold Coast and Slave Coast. Chapter XIV, "Folk-Lore Tales," runs pp. 243-274 and is organised in two parts: four tales that are not about the tortoise, then a separately introduced set of six "Tortoise Stories" at pp. 258-274 1.

Ellis introduces the tortoise set with a statement that is itself a piece of documentation:

We now come to those tales which may be called "Tortoise Stories," since the tortoise (awon) always plays a leading part in them. The tortoise has, in these tales, various superhuman powers attributed to him, and, in most, is described as acting craftily or mischievously 1.

He then makes the comparison that is still standard: the tortoise fills in the tales of the Slave Coast the place that the spider anansi fills in the tales of the Gold Coast 1. yoruba-alo records the same comparison from modern scholarship.

What is usable, and why

Four things make the tale texts worth using despite the framing.

They are given at length. Ellis prints continuous narrative with dialogue, not plot summaries. The reader gets what the characters said, in what order, which is the raw material a corpus needs and which most nineteenth-century sources do not supply.

The tales carry their opening formula. Every tale begins "My alo is about..." and Ellis separately records the teller's title, akpalo, his professional variant, akpalo kpatita, and the fact that he uses a drum 1. That is performance information, not just text. alo-the-opening-formula-and-the-teller treats it.

The footnotes preserve language. Ellis glosses Yorùbá words as they arise: ekuru and the proverb about the tedious visitor, oka, kekere-ke and its derivation from keke, kayin as ka-iyin, iwin, ado, iyale and iya-iwo, the drums bata, dundun and gangan. These glosses are the only Yorùbá in most of the tales and they are frequently the only thing that makes a plot point intelligible.

He records his own analytic observations as observations. The greed footnote at p. 274, quoted in full in alo-ijapa-and-the-rock-of-yams, is a generalisation Ellis draws across his own six tales and marks as his own reasoning. So is the la tortue / le tortu conjecture discussed in alo-buje-the-slender, which he explicitly hedges. A collector who separates what he was told from what he concluded is more useful than one who does not.

What has to be refused

The evolutionary hierarchy. Ellis's concluding chapter states his project directly: he sought to determine "to what extent different conditions of culture led to the modifications of religious conceptions," and he ranks his three volumes' subjects as "three stages of progress, the Tshi being in the lowest stage and the Yoruba in the highest," reasoning that the Yorùbá "were once in the social and mental condition in which the Tshi tribes are now" 1. This is nineteenth-century cultural evolutionism. It is the organising frame of the book and it is not adopted anywhere in this corpus. Nothing in the tale texts depends on it.

Generalisations about "the native." Ellis footnotes Adun's evasive answer to the Head as "the ordinary excuse of a native who has no pretext ready" 1. Where such phrasing appears in a footnote this corpus is quoting, it is quoted as an artifact of the source and labelled, as it is in alo-adun-and-the-borrowed-body.

Etymologies presented without evidence. Ellis derives ajapa from aja, elf, plus pa, to be bald or bare, and translates it "bald-headed elf" 1. The corpus reports this as Ellis's own analysis and does not endorse it: it is printed without tone marks by a non-specialist, and the standard modern forms of the name and its variants are recorded in owe-tortoise-ijapa-trickster from the proverb collections. The same caution applies to "broken-back," which he offers with "it probably refers to."

Totemism as an explanatory frame. Ellis speculates that totemism may lie at the root of the tortoise's prominence, then argues himself out of it in favour of the animal's remarkable qualities 1. The speculation is his, in the anthropological idiom of 1894, and the corpus records it as a historical position rather than a finding.

What is missing from the record

Ellis names no informants. He gives no dates, no places within Yorùbáland, no ages, no indication of who told him these tales or in what circumstances, and no statement of how they were transcribed or translated. He gives no Yorùbá text for any tale. Whether he worked through an interpreter is not stated.

This is the sharpest contrast with the other collectors used in this section. Parkinson, writing fifteen years later, states his tales came mostly from Ọ̀yọ́ people and were kept in the phrasing of a Yorùbá interpreter, which is two facts more than Ellis provides (alo-parkinson-1909-as-a-source). Omolaye, writing in 2023, tabulates each informant's name, age, sex, town and date of collection (alo-omolaye-2023-as-a-source). Against that standard Ellis's tales are anonymous, and every file in this section drawing on him is marked confidence medium for that reason.

There is one further gap worth naming. Ellis records an opening formula, "My alo is about...," and never mentions the Ààlọ́ o / Ààlọ̀ call-and-response that yoruba-alo documents as the genre's signature. Whether his tellers did not use it, or used it and he did not think it worth recording, or recorded only what followed it, cannot be determined from the text. The corpus notes the absence rather than explaining it away.

The single most valuable thing in the book

Ellis's greed footnote is, in this compiler's assessment, the most useful analytic sentence in the folklore chapter, because it is a generalisation drawn from a body of material by someone who had the whole body in front of him:

It is to be noted that, whenever he is shown as being over-reached or unsuccessful, the want of success is due to greediness 1.

He then enumerates his evidence across three of his six tales. Modern scholarship reaches the same conclusion, as recorded in yoruba-alo, and the corroboration from an 1894 corpus is worth having. It is also a reminder that the tale texts and the framing chapters can be assessed separately: a collector can be wrong about the evolution of religion and right about what his own tales do.

Where the tales are treated

Tortoise Stories I to VI: alo-ijapa-singing-drum, alo-ijapa-and-the-dumb-princess, alo-ijapa-rides-the-elephant, alo-adun-and-the-borrowed-body, alo-buje-the-slender, alo-ijapa-and-the-rock-of-yams.

The four non-tortoise tales: alo-the-palm-oil-seller-and-the-goblin, alo-the-poor-woman-and-the-aranran-bird, alo-why-the-ajao-remained-unburied, alo-the-king-and-kini-kini.