Seen with one’s own eyes: what Yoruba usage teaches about knowing and believing

When my grandmother wanted to know whether a piece of news was reliable, she did not ask, "Is it true?" She asked who had seen it. The difference between what a person has witnessed and what they have been told runs through Yoruba speech, and in my doctoral work I have spent several years thinking about what it means for epistemology. This essay sets out the distinction, places it beside the main positions in the philosophy of testimony, and argues that it captures something those positions tend to lose.

The distinction

In Yoruba the verb mọ̀ is used for knowing, and the noun ìmọ̀ for knowledge. The verb gbàgbọ́, which can be analysed as accepting what one hears, gives us ìgbàgbọ́, belief or what one has accepted. Over a number of years two philosophers, Barry Hallen and J. O. Sodipo, held long conversations with Yoruba oníṣègùn, specialists in herbal medicine, about how these words are used. Their book Knowledge, Belief, and Witchcraft (1986) reported a pattern that surprised many readers. Among their interlocutors, mọ̀ was properly used for what one had experienced directly, above all what one had seen with one's own eyes. What one had been told, however reliable the teller, belonged to ìgbàgbọ́: it was something one had accepted, and it remained open to being tested.

The consequence is striking. Much of what an English speaker would call knowledge, the date of a battle, the capital of a distant country, most of what is learned from books and teachers, would on this usage be belief: properly accepted, perhaps, but not known in the strict sense.

I should add a caution, because careful scholarship requires it. Hallen and Sodipo worked with a particular group of specialists, and some readers have asked whether their reflective usage represents ordinary Yoruba speech, or whether the account makes the distinction sharper than everyday life does. That discussion continues. But the contrast itself, between first-hand witness and received report, is familiar to any Yoruba speaker, and that is enough to make it philosophically interesting.

The debate it enters

In contemporary epistemology the question of testimony is roughly this: when are you entitled to believe something because someone told you?

Reductionists answer that testimony is justified only through other sources. You may believe a speaker because you have evidence, from your own perception and inference, that this kind of speaker on this kind of topic is usually right. David Hume's discussion of miracles weighs testimony in this way, against our experience of how often such reports turn out true.

Anti-reductionists answer that trust is the rational default. You do not need positive reasons to believe what you are told; you need reasons to doubt it. Thomas Reid argued that human beings have a natural disposition to believe others, matched by a natural disposition to speak the truth, and that a creature which believed nothing until it had checked the speaker's record would never learn to speak at all.

Classical Indian philosophy drew a parallel line. The Nyāya school treated testimony from a reliable speaker as an independent means of knowledge, while Buddhist logicians reduced it to inference.

More recently, Miranda Fricker has shown that the default of trust is not distributed fairly. Credibility is often withheld from speakers because of who they are, a wrong she calls testimonial injustice, which harms the speaker and also leaves the hearer with less of the truth.

What the Yoruba distinction adds

At first glance the Yoruba picture looks reductionist: testimony yields accepted belief, not knowledge. But that misreads it. It does not say we are unjustified in accepting what we are told. Ìgbàgbọ́ is not a term of reproach. It says that what we accept on someone's word has a different epistemic standing from what we have witnessed, and that the difference should be kept in view.

I want to argue that this is an advance on both sides of the Western debate.

The anti-reductionist is right that we must trust in order to know much at all. But anti-reductionism tends to blur the line between a belief I have tested and a belief I have received, treating both as knowledge once the default is granted. Over time, as memory loses the source, the received belief starts to feel like something I saw. The Yoruba usage resists that blurring. It keeps a label on the belief that says, in effect, "accepted on report; open to verification".

The reductionist is right that testimony can be checked, but wrong to make checking a precondition of reasonable acceptance. The Yoruba usage allows acceptance without prior checking, while preserving the standing possibility of it.

There is a social side as well. If the line between ìmọ̀ and ìgbàgbọ́ stays visible, the witness has a distinctive role and responsibility: the one who saw. In a society where much important knowledge travelled orally, from elders, from specialists, through memorised verse, a clear marker of who actually witnessed what is a practical safeguard. It is not the mark of a "pre-scientific" mind, as some older anthropology suggested about African thought. It is a careful epistemic practice.

Why it matters now

Kwasi Wiredu urged African philosophers to test philosophical concepts by trying to think them through in their own languages, to find out which problems survive translation and which are artefacts of English. The knowledge and belief distinction is a good test case. It suggests that the English word "knowledge" stretches over two different things: what we have seen and what we have been told.

That distinction has never been more useful. Every day I receive messages, forwarded and reforwarded, carrying claims that their senders sincerely believe and have never witnessed. A culture of reasoning that asked by habit, "Did you see it, or were you told?", and that treated being told as respectable but provisional, would be better equipped for such a world.

My conclusion is modest but, I think, firm. The Yoruba distinction is not a curiosity for comparative tables. It is a position in the epistemology of testimony: trust is permitted, but testimony's status should stay marked, and the witness's word and the hearer's acceptance are different things. Western epistemology has the resources to state that position, but it has rarely placed it at the centre. My grandmother did.