Ewé: The Plant Knowledge

Ewé means leaf, and it stands for the entire botanical body of Yoruba medicine. The system's own summary of itself is the proverb ewé ni oògùn, leaf is medicine, and the corresponding claim in the Ifá literature that there is no plant which is not medicine. The documented scale is large: Pierre Verger's Ewé, the standard reference, records information on the order of three and a half thousand plants gathered over four decades of fieldwork, together with more than two thousand remedies, the plant parts used, the preparation methods, the incantations that accompany them and their attachment to Ifá divination signs 12.

This file does two things. It describes how the knowledge is organised, named and remembered, which is where the tradition is most distinctive and least understood. Then it sets a worked table of named plants against the pharmacological literature, marking for each one exactly what has been demonstrated, in what model, and what has not.

A warning about the table before the table: the honest answer for most of the pharmacopoeia is that nobody has looked. That is not a verdict on the tradition. It is a statement about where research funding has gone.

Verger's corpus

Pierre Fatumbi Verger (1902-1996) was a French photographer and ethnographer who settled in Bahia, was initiated as a babaláwo, took the name Fatumbi, and spent decades moving between Brazil and Yorubaland. Ewé: The Use of Plants in Yoruba Society was published in 1995, runs to 744 pages, and is the single largest published record of Yoruba plant knowledge 1.

What makes it unusual is not the species list. It is that Verger recorded the plants together with the ọ̀fọ̀, the incantations, and with the odù Ifá to which the formulae attach 2. Most ethnobotany records a plant, a use and a preparation. Verger recorded the plant inside the verbal and divinatory apparatus that a practitioner actually uses it within, which is why the book is a source for the structure of the knowledge and not only for its content.

Two cautions about using it.

It is a record of what practitioners said, not a pharmacological evaluation. Verger documented; he did not test. A use recorded in Ewé is evidence that the use is traditional, which is a real and valuable thing to establish, and it is not evidence that the use works.

Its publication is not uncontroversial among practitioners. Much of the material was held under obligations of secrecy, and putting it in print is regarded by some as a breach and by others as a rescue. Both positions are held by initiated people and this compiler does not adjudicate. The tension is the same one that runs through all publication of initiate-held knowledge and it recurs in 05-ifa.

How plants are classified

There is more than one classification in use and they operate at different levels. A reader should not expect a single Linnaean-style hierarchy, because that is not what the system is for.

By element. A documented scheme sorts leaves into four groups: ewé afẹ́fẹ́, leaves of air or wind; ewé iná, leaves of fire; ewé omi, leaves of water; and ewé ilẹ̀ or ewé igbó, leaves of the ground or of the forest 3. Schemes vary by source and by practitioner lineage, and 06-orisa/11-osanyin treats this alongside Ọ̀sanyìn, the òrìṣà who holds the knowledge of leaves.

By sensory property. Working herbalists identify and group plants by odour, by leaf texture, by how the plant reacts when touched and by the sensation felt on contact. These properties enter into naming as well as into classification, which is the link to the next section.

By action. The functional grouping that matters most in practice: what a plant does. Bitterness in particular is a recognised class marker, and it maps onto the kòkòrò and aràn framework described in file 01, where bitter-tasting plants are the characteristic agents for controlling the organisms held responsible for disease.

By òrìṣà. Plants attach to deities, and a preparation for a condition governed by a particular òrìṣà draws on that òrìṣà's plants.

These cut across one another. A plant has an element, a set of sensory properties, a set of actions and a set of divine associations at once, and a practitioner uses whichever axis the problem calls for. That is a working classification rather than a taxonomy, and judging it against Linnaeus mistakes what it is doing.

The name as mnemonic: the wordplay that carries the pharmacopoeia

This is the most distinctive feature of Yoruba pharmacological memory and it is routinely either ignored or mystified. The finding is Buckley's, from fieldwork with herbalists: most medicinal incantations use a form of word-play, similar to punning, to evoke the properties of the plants implied by the name of the plant 4. Verger's material is organised around the same relationship 2.

Here is what that means mechanically. Yoruba is a tonal language with heavy homophony, and a great many plant names are transparently compound or descriptive. An incantation over a preparation takes the plant's name and puns on it, so that the name is made to say what the medicine is supposed to do. Some worked examples of the naming logic, at the level the published sources support:

Ewé ìṣẹ́nbáyé. The name of a plant is frequently a compressed sentence. Names of the form a-...-bá-... or built on verbs of doing and arriving are common, and the incantation extracts the verb.

Dòkítà igbó, forest doctor, is a Yoruba name for Enantia chlorantha 5. That name is not a mnemonic pun but a straightforward claim about the plant's standing, and it tells you the plant is important before any chemistry does.

Ewúro, Vernonia amygdalina, bitter leaf. The name marks the property that classifies it.

Dòngòyárò for Azadirachta indica, neem. The name is a loan reflecting the plant's arrival in West Africa in the colonial period, and it is a useful reminder that the pharmacopoeia is not a closed ancient inventory. Neem is a nineteenth and twentieth century introduction that was absorbed into Yoruba practice, named, and given uses. Traditions take in new material.

Two things follow from the naming system and both matter.

It is a memory technology. In a system without writing, tying a plant's use to its name means the practitioner who has the name has a hook to the use. The incantation is the retrieval mechanism as well as the activating agent. This is a sophisticated solution to the problem of holding several thousand plant-use pairs in memory, and it is the reason the pharmacopoeia is as large as it is.

It creates a systematic risk of false positives. If a plant's use is derived from its name by punning, then some uses will have been generated by the name rather than discovered by observation. The mechanism that makes the knowledge memorable also manufactures plausible-sounding indications with no empirical basis behind them. This is the single most important thing to understand about the epistemic status of the pharmacopoeia, and it cuts precisely the way the evidence in the table below suggests: some plants have real activity that practitioners identified without any laboratory, and some have indications that are artefacts of Yoruba phonology. Only testing separates the two, and testing is exactly what has mostly not been done.

Stating this plainly is not a dismissal. Signature-based reasoning generated a great many indications in European herbalism too, and European herbalism also produced digitalis and salicylate. The point is that the presence of a use in the tradition is weak evidence of efficacy by itself, and strong evidence only that the use is traditional.

The scale of the pharmacopoeia

Verger's roughly 3,549 plants and 2,000-plus remedies is the largest figure 12. Contemporary ethnobotanical surveys record smaller local pharmacopoeias, which is what you would expect since they sample a town or a district rather than the whole tradition: a survey of Ile-Ife documented 87 species in 43 families from 70 informants 6. Larger inventories of Yoruba useful plants across southwestern Nigeria have been published in Heliyon 7.

Two features of the survey literature are worth reporting because they bear on whether this is systematic knowledge or individual improvisation.

Agreement between independent practitioners is high for the commonly used species. Where quantitative indices are calculated, a small number of species recur across informants with high citation frequency, which is what transmitted systematic knowledge looks like and not what individual invention looks like.

Leaves dominate. Across surveys the leaf is by far the most used plant part, which is consistent with ewé standing for the whole field, and herbs and shrubs predominate over trees in the counts.

A research-integrity caution. One frequently cited quantitative ethnobotany of Ile-Ife, published in Evidence-Based Complementary and Alternative Medicine in 2021, was subsequently retracted by the publisher following an investigation that found evidence of systematic manipulation of the publication and peer-review process 6. Its numbers still circulate widely in later citing literature. This corpus reports its figures only with the retraction attached, and a reader encountering Nigerian ethnobotany citations should be aware that the Hindawi retraction wave of 2023 removed a significant number of papers in this field.

The table

What follows is a worked table of named plants. The evidence column is graded deliberately, and the grades mean specific things:

• Human trial evidence means at least one controlled trial in people, with the result stated whether positive, negative or inconclusive.
• Animal or in vitro only means activity has been demonstrated in a model, with the compound named where one has been isolated. This does not establish clinical efficacy.
• Compound isolated, activity uncharacterised means chemistry has been done but therapeutic activity has not been established.
• Traditional use only means no pharmacological evaluation was located for the specific traditional indication. This is the commonest category in the pharmacopoeia as a whole and it is under-represented in this table, because the table selects for plants that have been studied.

The table is not a therapeutic guide and nothing in it should be read as a recommendation. Doses in traditional use are not standardised, extracts in studies are not the same material as traditional preparations, and several of these plants are toxic at achievable doses.

Yoruba name	Botanical binomial	Traditional use	What research has established	Grade
Dòkítà igbó, awópa	Enantia chlorantha Oliv. (Annonaceae)	Malaria, fever, jaundice; one of the most-cited Yoruba antimalarials	Protoberberine alkaloids present; the alkaloid jatrorrhizine reported with ED50 0.34 mg/g ethanolic and 6.9 mg/g aqueous extract against Plasmodium yoelii 8. Toxicological study of the purified protoberberine alkaloidal fraction in 120 mice found no pathological lesions in stomach, kidney, oesophagus or liver at the doses tested, mild to moderate lung oedema, and no significant biochemical differences from control, suggesting relative safety of short-term use 9	Animal and in vitro only
Òrúwo	Morinda lucida Benth. (Rubiaceae)	Malaria, fever, jaundice	Bioassay-guided isolation identified antimalarial triterpenoid acids including asperulosidic acid and asperuloside; parasitaemia reduction of 51.52 percent against P. berghei NK65 and chemosuppression in the range 39.8 to 90.5 percent reported 8. A 2023 study of leaf combinations reported 90.7 to 91.0 percent chemosuppression at 16 mg/kg for two-plant combinations against chloroquine's 100 percent at 10 mg/kg, with oleanolic acid and ursolic acid identified by molecular docking as multiple-target inhibitors 10	Animal and in vitro only
Awùn, àhun	Alstonia boonei De Wild. (Apocynaceae)	Malaria, fever, pain, snakebite	Reported chemosuppression ranging from 0.2 to 74.8 percent in P. berghei infected mice 8. The width of that range across studies is itself the finding: the material is not reliably potent under the conditions tested	Animal and in vitro only, inconsistent
Dòngòyárò	Azadirachta indica A. Juss. (Meliaceae)	Malaria, fever, skin conditions; a colonial-era introduction absorbed into the pharmacopoeia	More than 400 compounds isolated from the plant including azadirachtin, nimbidin, nimbin, nimbolide and gedunin 11. Gedunin identified as an antiplasmodial constituent, with dose-dependent suppression of 69.65 to 78.32 percent reported 8; bioassay-guided work on the fruits confirmed significant in vitro antiplasmodial activity for gedunin and azadirone and identified two new triterpenoids, neemfruitins A and B 12	Animal and in vitro only
Kóóko oyinbo	Cymbopogon citratus (DC.) Stapf (Poaceae)	Fever, malaria, taken as an infusion	Geranial identified in the essential oil; IC50 of 4.2 micrograms per millilitre against P. falciparum reported 8	In vitro only
Ewúro	Vernonia amygdalina Delile (Asteraceae)	Malaria, fever, stomach complaints, diabetes; also a food	Extensive published antimalarial and antidiabetic screening in animal and in vitro models. This compiler did not locate a controlled human trial for the malaria indication	Animal and in vitro only
Orógbó	Garcinia kola Heckel (Clusiaceae)	Chewed for cough, chest and throat complaints, as a stimulant, for liver conditions	The biflavonoid complex kolaviron, comprising GB1, GB2 and kolaflavanone, is isolated and characterised; also garcinoic acid, garcinal, kolanone, gakolanone 13. Hepatoprotective, cardioprotective, neuroprotective, anti-inflammatory, antidiabetic and antimalarial activity documented in rodent models 13. A randomised human trial in knee osteoarthritis exists 14. The 2023 critical review states that clinical data in humans on any constituent of G. kola are entirely missing, that most studies use doses above 100 to 400 mg/kg which are clinically unrealistic and correspond to roughly 7 to 14 g of pure substance in a human, that very few studies used appropriate controls, and that positive controls where used were often applied at around one hundredth of the kolaviron dose 13	Animal models plus one human trial; the review literature judges the human evidence base absent
Ẹfinrin, ẹfinrin nla	Ocimum gratissimum L. (Lamiaceae)	Fever, diarrhoea, skin infections, infertility, taken as leaf decoction	Eugenol is the major essential-oil constituent, reported at 74.83 percent in one analysis; thymol, carvacrol, limonene, beta-caryophyllene, rosmarinic acid, caffeic acid, quercetin and luteolin also identified 15. Antioxidant, anti-inflammatory, antimicrobial, antidiabetic, hepatoprotective, analgesic and neuroprotective activity documented in animal and in vitro work. The 2021 review states explicitly that further human clinical trials are needed to establish effective and safe doses, and identifies no human trials 15	Animal and in vitro only
Ẹ̀gbọ̀, ìbẹ̀pẹ	Carica papaya L. (Caricaceae)	Leaf extract for fever and, in modern popular use across the tropics, for low platelets in dengue	Nine controlled trials meta-analysed. Seven showed increased platelet counts; mean improvement in platelet count between days one and five of 35.45 (95% CI 23.74 to 47.15, three studies, 129 participants, graded low quality evidence); possible reduction in hospital stay of 1.98 days (95% CI 1.83 to 2.12, three studies, 580 participants, low quality). The authors conclude current evidence is insufficient to comment on the role of the extract in dengue, because the clinical value of a platelet rise without harder outcome measures is unclear 16	Human trial evidence, low quality, inconclusive
Ejinrin wéwé	Momordica charantia L. (Cucurbitaceae)	Diabetes, fever, worms	Randomised placebo-controlled trial in type 2 diabetes, 90 subjects analysed, 12 weeks: HbA1c unchanged in both groups; mean fasting glucose decreased in the extract group (p = 0.014); no serious adverse events 17. So: a measurable effect on fasting glucose, no effect on the standard measure of glycaemic control	Human trial evidence, mixed
Àsọ̀fẹ́yẹjẹ, ìra	Rauvolfia vomitoria Afzel. (Apocynaceae)	Mental disturbance, agitation, hypertension, snakebite	The genus is the source of the monoterpene indole alkaloids reserpine, ajmaline, ajmalicine, serpentine and yohimbine 18. Reserpine was developed into a mainstream antihypertensive and early antipsychotic, and clinical trials exist for R. serpentina formulations and for the alkaloids themselves; for R. vomitoria specifically the review notes the extracts and constituents await clinical investigation 18. This is the clearest case in the Yoruba pharmacopoeia where the traditional indication, agitation and high blood pressure, matches the pharmacology of an isolated compound that entered world medicine	Compound isolated and clinically developed as a single agent; the plant as traditionally used is not itself trialled
Èlú	Philenoptera cyanescens (Schumach. & Thonn.) Roberty, formerly Lonchocarpus cyanescens (Fabaceae)	Indigo dye; also traditional use for skin conditions and as a wash	Chemistry of the dye is well characterised: indigotin plus 2-hydroxynaphthoquinone and mineral ions. The dyeing chemistry is treated in file 09 and 07-arts/05-textiles. Medicinal indications not evaluated in any study located	Traditional use only for the medicinal indication
Tagìrì	Adenopus breviflorus Benth., now Lagenaria breviflora (Cucurbitaceae)	Placed around the house to stop the spread of smallpox and measles; a prophylactic rather than a treatment 19	No study located evaluating the prophylactic claim. As a physical barrier to an airborne virus the claim has no plausible mechanism	Traditional use only
Ewé àbámọ̀dá	Bryophyllum pinnatum (Lam.) Oken (Crassulaceae)	Applied to wounds and boils, taken for various complaints	Bufadienolides isolated from the genus and characterised; these are cardiac-active steroids and the toxicological significance is real	Compound isolated; traditional indications not established
Ataare	Aframomum melegueta K.Schum. (Zingiberaceae)	Grains of paradise; ritual use in prayer and offering, also digestive and stimulant use	Substantial phytochemical and pharmacological literature in animal and in vitro models. No human trial located for the traditional indications	Animal and in vitro only
Àgbálùmọ̀	Chrysophyllum albidum G.Don (Sapotaceae)	Fruit eaten; bark and leaf used for various complaints	Antioxidant and related activity documented in laboratory models	Animal and in vitro only
Èèrù	Xylopia aethiopica (Dunal) A.Rich. (Annonaceae)	Post-partum preparations, digestive complaints, spice	Pharmacological screening published in animal and in vitro models	Animal and in vitro only
Egúsí, ẹ̀wọ̀n	Nauclea latifolia Sm., now Sarcocephalus latifolius (Rubiaceae)	Malaria, fever, pain	Indole alkaloids isolated from the species; the genus is among those to which most bioguided attention has been paid within sub-Saharan Rubiaceae, with antimalarial, antimicrobial, antihypertensive, antidiabetic, antioxidant and anti-inflammatory activities screened 20	Animal and in vitro only

Reading the table honestly

Four conclusions follow and they should be held together.

The identifications are frequently real. Practitioners picked out plants that contain compounds with measurable activity, and they did it without microscopes, culture, animal models or chemistry. That is a genuine epistemic achievement and it deserves to be said without hedging. The antimalarial cluster is the strongest instance, and its plausibility is underwritten by the two most important antimalarials in world medicine, quinine from Cinchona and artemisinin from Artemisia annua, both of which came out of exactly this kind of tradition 8.

Demonstrated activity in a model is not demonstrated efficacy in a person. Almost all of the evidence above is in vitro or in rodents. Extracts are not standardised, potency varies enormously with plant material and extraction method, and the doses that produce effects in animals frequently correspond to human doses nobody would take. The Garcinia kola review makes this point with unusual force and precision, and its criticism generalises across the field 13.

Where human trials exist, the results are modest and mixed. Carica papaya raises platelet counts in dengue on low-quality evidence with no demonstrated effect on outcomes that matter 16. Momordica charantia lowers fasting glucose but not HbA1c 17. This is what an honest reading of the human evidence looks like, and it is neither a vindication nor a refutation.

Absence of evidence is mostly absence of research. For the large majority of the several thousand plants Verger recorded, and for most indications other than malaria, no pharmacological evaluation exists at all. The research effort has concentrated where the funding is. A reader should not infer from the malaria results that the tradition is generally validated, and should not infer from the silence elsewhere that it is generally wrong. Most of it has simply never been examined.

The corresponding safety point belongs here too: these plants are not inert, which is the same property that makes them potentially useful. Abortifacient effects have been examined in commonly used African antimalarial plants in pregnant mice 21, and the contamination and quality problems in the contemporary trade are treated in file 12.