Joel at Far Outliers is posting excerpts from Britain at War with the Asante Nation, 1823–1900: “The White Man’s Grave” by Stephen Manning (Pen & Sword Books, 2021), and the post Foundation of the Asante Nation includes this etymology:

The name Asante seems to have derived from a special red clay the people sent to the dominant tribe, the Denkyira, as a form of payment or tribute of allegiance. The Akans call clay ‘Asan’, therefore the Asantes were differentiated from others with the name ‘Asan-tefo’, or those who dig clay.

Of course this aroused my curiosity, so I googled and discovered there is an alternative version, presented by Wikipedia thus:

The name Asante means “because of war”. The word derives from the Twi words ɔsa meaning “war” and nti meaning “because of”. This name comes from the Asante’s origin as a kingdom created to fight the Denkyira kingdom.

That sounds very much like a folk etymology to me, but of course the same could be true of Manning’s; anybody know more about this? (Wiktionary says “From Twi asànté,” which isn’t much help.)

Incidentally, the OED doesn’t have an entry for Asante, though it has one (new as of June 2022) for the phrase asante sana (East African) ‘thank you very much’:

Etymology: < Swahili asante sana < asante, ahsante thank you (< Omani Arabic aḥsante thank you: see note) + sana (adverb) very, much.
Omani Arabic aḥsante thank you (19th cent. or earlier) is a specific sense development of aḥsante ‘well done! bravo!’ (literary Arabic aḥsanta), use as interjection of the 2nd person masculine singular of aḥsana to do (something) well, to perform (something) nicely < the same base as ḥusn beauty. Due to the long historical and trade links between Oman and the East African coast and Zanzibar, Omani Arabic was a major source of Swahili loans from Arabic.

Which is great — I’m glad they have the space to include such details of word history, and the breadth of vision to include regional terms like this — but it points up the foolishness of not including a far more common word like Asante simply because it refers to a nation.


  1. I don’t have access, but it seems Ashanti is in OED.

  2. Dan Milton says

    OED: “Etymology:Self-designation”.
    Not much help there!

  3. < Omani Arabic aḥsante thank you: see note


    (19th cent. or earlier) – I wonder what this means.

  4. Gedzi, Victor Selorme. “The Asante of Ghana.” International Journal of African Society Cultures and Traditions 2.3 (2014): 20-26.

    Who are the Asante? It is therefore appropriate that in this initial stage we define the name Asante (Asanteni – an Asante; Asantefo – Asante). One may have to start with Asantefo as a word and as a linguistic problem. What did the word Asantefo originally mean? What is its etymology? Historians and anthropologists continue to debate this issue and have offered various answers (see Allman 2000; McCaskie 1995; Osei 2001; Wilks 1993). This paper addresses three main attempts made to arrive at a Twi etymology for Asante. These are namely,

    * Asan-te-fuo (clay producing group),

    * Esa-nti-fuo (because of war group) and

    * Asanteni-ba (a child of Asante). [I think Gedzi is confusing a recursive definition of ethnicity with a recursive etymology.]

    None of the references look like linguists:

    Allman, J.M. (2000) ‘Be(com)ing Asante, Be(com)ing Akan: Thoughts on Gender, Identity and the Colonial Encounter’, in Carol Lentz and Paul Nugent (eds), Ethnicity in Ghana. The Limits of Invention. pp. 97-118 New York: St. Martin’s Press.

    McCaskie, T. C. (1995) State and Society in Pre-colonial Asante. Cambridge: Cambridge University Press.

    Osei, K. (2001) An Outline of Asante History. Part 1 (3rd ed.). Suame-Kumasi: O. Kwadwo Enterprise

    Wilks, Ivor (1993) Forests of Gold: Essays on the Akan and the Kingdom of Asante. Athens: Ohio University Press.

  5. David Eddyshaw says

    It’s true that ɔsa means “war” and nti is a postposition meaning “because of”*, but the Twi for “Ashanti” is Asante, not *Asanti. The tone of the first supposed component differs, too, but I think that can happen in Twi compounds. Still, it looks like a classic folk etymology to me, too.

    Twi rather lends itself to folk etymology with all its very short CV stems. In Christaller’s dictionary, there are sa stems meaning “cut”, “draw”, “scoop”, “dance”, “bring into”, “come to an end”, “path”, “war”, “largest room in the house”, “loom”, “warp”, “hand”, “agree”, “mix” and “saw”, and stems “san” meaning “draw a line”, “infect”, “loosen” and “shed”: anybody who can’t make up several folk etymologies with such abundant materials should forswear that art altogether.

    The alleged asan “clay” is not in Christaller’s (very comprehensive) Twi dictionary at all, though, Seems to be a figment.

    Christaller says that the original form of “Asante” was actually Asiante; he doesn’t particularise, but I think this is based on comparison with the Fante and Akuapem forms.

    Most ethnonyms in the Oti-Volta zone are unanalysable (Batammariba, “builders” is very much the exception to prove the rule.) I don’t know much about Akan onomastics, though. It may be different.

    The usual word for “Ashanti” in northern Ghana is reflected in the Kusaal Akambʋŋ “Ashanti person.” I suppose that the first component may be something to do with “Akan”; Kusaasi happily tell you that the second component is bʋŋ “donkey”, but I’m pretty sure that that’s folk-etymology too …

    * Presumably related to eti “head”: Western Oti-Volta likewise uses “head” (Kusaal zug) as a postposition “upon”, very frequently with the metaphorical extension to “on account of.”

  6. it seems Ashanti is in OED.

    Ouch, I should have thought to try that! But, as Dan points out, it’s no help in this regard.

  7. * Asanteni-ba (a child of Asante). [I think Gedzi is confusing a recursive definition of ethnicity with a recursive etymology.]

    Yes, that’s very odd.

  8. On the subject of odd etymologies: back in 2008, I blogged a bit about two works on how body-part terms yielded prepositions in Oceanic (John Bowden) and African (Bernd Heine) languages. I’ve added two snippets below.

    “Heine examined sources for African locatives in five categories, ON, UNDER, IN, FRONT, BACK, to which Bowden added SEA, LAND, and OUT for Oceanic languages. In both Africa and Oceania, body-part nouns provide the most common sources for locatives in the categories ON, IN, FRONT, and BACK, whereas landmark nouns (‘earth’, ‘soil’, ‘shadow’) predominate for UNDER. As might be expected, landmark nouns also predominate for SEA and LAND locatives in Oceania. The exceptional cases are instructive. For instance, in both Africa and Oceania, ‘head’ is the most common source (‘sky’ is next) for ON, while ‘face’ is the most common source for FRONT. But among some quadruped-herding populations in Africa, FRONT derives from ‘head’, while ON derives from ‘back’.”

    “Bowden alludes to a more interesting development in a footnote: locatives that come full cycle and yield euphemistic terms for body parts (‘down below’, ‘inside’, ‘backside’, etc.).”

  9. [I think Gedzi is confusing a recursive definition of ethnicity with a recursive etymology.]

    A cpossible meaningful reading is: “doesn’t contain any formants characteristic for ethnonyms, because it was a name of something else”.

  10. Might Asante share a morpheme with Fante?

    (WP: ‘According to oral tradition the Fante separated from the other Akan groups in present day Brong Ahafo around 1250 AD. This act became the origin of their name, “Fa-atsew” meaning “half that left”.’ I don’t know…)

  11. “Asante” seems to be taking over as the spelling that the OED uses in editorial matter; it appears in the revised etymologies of ackee, Akani, and balm n.2, and in the definition of Akan (2012): “2. The group of closely related Niger-Congo languages spoken by these peoples, comprising Asante, Fante, and Twi; any one of these languages.” However, the spelling “Ashanti” still survives in the revised definition of messenger sword (s.v. messenger) and etymology of adinkra (cloth pattern), and in the etymologies of unrevised kankie (type of bread) and sancho (musical instrument). It seems likely they’ll change the primary spelling of the headword when it gets revised.

  12. David Eddyshaw says

    comprising Asante, Fante, and Twi

    An extremely odd way of putting it. The Asante/Ashanti speak (one kind of) Twi, the other major literary form of which is Akuapem (Twi): the term “Twi” is generally used to mean essentially “not-Fanti”, and is quite often loosely used for the whole dialect cluster. But on any showing, “Asante and Twi” is just wrong. It’s like saying “Canadian English and English.”

    I tend to slip into using “Twi” to mean “Ashanti Twi” myself, which is strictly speaking not quite accurate; though the Ashanti are surely (in Sellar/Yeatman terms) the most memorable of the Akan speakers.

    The term “Akan” is itself not ideal: there are culturally Akan peoples who speak related languages which are nevertheless quite distinct from the Ashanti/Akuapem/Fante language*; but it’s long since been canonised in Ghana as the accepted name for that particular dialect cluster. It was invented in relatively recent times, precisely because there was previously no name for the whole language.

    * Kwame Nkrumah’s mother-tongue, Nzema, is a case in point.

  13. David Eddyshaw says

    body-part nouns provide the most common sources for locatives

    Certainly true for “head/upon” in Oti-Volta, and also “belly/inside”, but “before” and “behind” don’t seem to be based on body parts synchronically or historically.

    Kusaal distinguishes teŋin “downwards”, which is transparently just the locative of teŋ “ground”, from teŋir “below”, which is evidently related, but not by any known derivational process.

    Gbin “buttock” is what hills and mountains have, rather than “feet”, as in English. Kusaasi hills are presumably conceptualised as lazier than their Brit counterparts.

    Gbin has ended up also meaning “meaning”, as in Luke 18:34 ka ba pʋ baŋ on yɛt si’el la gbinnɛ “but they did not understand the buttock of what he was saying.”

  14. Stu Clayton says

    “but they did not understand the buttock of what he was saying.”

    Buttocks are the seat of reason. Makes sense.

  15. In Numbami, dabola ‘head, base, stump’ can also mean ‘basis, reason’. Also tinadaba ‘headwater’. There is a local legend about an offshore island that picked itself up and moved to where the port city of Lae now sits because the villagers were not feeding it as much as its neighboring mountain, but I’m not sure whether the mountain moved by flying, swimming, or wading.

    Every body part that has a hole also ends in -owa, from awa ‘hole’: tanganowa ‘ear’, nisinowa ‘nose’, tinowa ‘penis’, inowa ‘vagina’, tainowa ‘arse’, but the ending is no longer analyzed, so to say, ‘nostril’, for instance, you have to add awa: nisinowa awa. And you differentiate inner from outer ear by tanganowa awa vs. tanganowa lau (lau ‘leaf’).

  16. ktschwarz says

    “Asante and Twi” is just wrong

    Thanks, I didn’t realize that. Disappointing for a 2012 OED entry. Let’s see if AHD does better:

    Ashanti¹ 2. The Twi language of the Ashanti.
    Twi A variety of the Akan language spoken in Ghana.
    Fante or Fanti 2. The variety of Akan spoken by this people.
    Akan 1. A Kwa language spoken in parts of Ghana and Côte d’Ivoire whose two main varieties are Fante and Twi.
    Kwa Any of several West African languages belonging to the Niger-Congo language family, including Ewe.

    That’s correct, isn’t it? Except they probably should have called Kwa “a group of languages” rather than “any of several”?

  17. David Eddyshaw says


    Kusaal just uses its inherited Oti-Volta birthright of freely forming compound nouns in such cases: gbinvɔɔnr “buttock-hole” …

    I’m not sure whether the mountain moved by flying, swimming, or wading.

    In any case, it puts Brit mountains to shame, for merely standing about: all the more so, Kusaasi mountains, which can’t even be bothered to do that, and just sit around all day. Obviously the trick is not to overfeed them.


    Yes, I think that’s good enough (and certainly better than the OED): I agree about “Kwa”, where they’ve expressed themselves comprehensibly but sloppily.

    I might object to a Gbe language like Ewe being classified as “Kwa”, but I suppose one has to accept that Greenberg remains state-of-the-art as far as general reference works are concerned. (Inevitably so, I suppose, given that there is more agreement among the experts about where he was wrong in his subclassifications than there is consensus over what is actually right.)

  18. @ktschwarz: You mentioned the Asante messenger sword, and I Googled since I didn’t remember every having seen a picture of one. Most of the Google hits were for items in various fantasy games, but the there were some images of what I was searching for. The World Museum Liverpool has a messenger sword, although the original documentation of its provenance was apparently lost during the Blitz.

    Mention of kankie also reminded me of the delightful Asante folktale “Half-A-Ball-Of-Kankie.” I read an illustrated version of it as a child, and I thought it was hilarious. The title is usually hyphenated in English, because it is actually the name of the anthropomorphic hero of the tale.

  19. When I was 17, a friend of mine boasted that he is learning Swahili and drew on my wall some hybrid between a human face and a ritual mask (not very different from his usual highly ornamental drawing style) and a phrase asanteni sana mwendawazimu.

    From his translation (ну типа “спасибо, придурки”) and m- I assume he did not know much about Swahili grammar:)

    (This is why “Aaaaa” above)

  20. David Eddyshaw says


    The original seems to have been collected (and presented along with a parallel Twi text) by no less a person than Tribes-of-the-Ashanti-Hinterland R S Rattray himself. (Sadly, I can’t find a downloadable version.)

  21. Jen in Edinburgh says

    Do we know what mountains eat?

  22. David Eddyshaw says


  23. If I weren’t too close to a wall, I’d flip backwards.

  24. Are the similar endings of Asante and Fante a coincidence?

  25. David Eddyshaw says

    Well, the tones match as well as the vowels. But as far as I can see, “Fante” doesn’t have any genuine recognisable etymology either, so the question seems unanswerable. There isn’t a general pattern of ethnonyms in -(n)te. (Hausa productively makes names of languages out of ethnic group names by adding -anci, e.g. Turanci “English/French”, Yarbanci “Yoruba”, Buzanci “Tamasheq”, but I can’t think of any even remotely plausible way that this could be connected.)

    Northern Ghanaian ethnic group names seem practically never to be analysable (unlike place and personal names. which are nearly all synchronically transparent.) That doesn’t prevent people coming up with folk etymologies, which are as popular in Ghana as elsewhere. The Kusaasi like to tell you that their name comes from the Hausa kusa “near”, an idea which seems to be based entirely on the similarity of sound and the fact that, if you actually are Kusaasi, “near” seems a reasonable enough description of the Kusaasi in general.

  26. Gbin has ended up also meaning “meaning”

    is that a result of the bible translation? and if so, is it a triangulation through english “fundament”?

  27. David Eddyshaw says

    is that a result of the bible translation?

    No, I’m sure it’s not.
    Gbin basically (sorry) means “base” (it’s used for the “foot” of a tree, too), and I suspect that the meaning “bum, backside” is actually secondary rather than primary (though it is the normal word for that essential part of the human body.)
    It may have started out as a euphemism. There are parallels; for example, Kusaal pɛ̀n “vagina” (again, the ordinary word, and with no other sense in Kusaal) is cognate to Mooré pèndé “lower abdomen.”

    I was going to try make a joke about “fundamentalism”, but couldn’t quite make it come right ….

    Interesting about the etymology of Swahili asante, BTW.

    Does anyone have any idea (talking of Swahili etymology) where mfalme “king” comes from? It can hardly be Bantu, but I’ve no clue where it might be loaned from.

  28. PlasticPaddy says

    gives origin from Bantu, but their source seems to be an 1899 Meinhof publication.

  29. David Eddyshaw says

    Yes; and it just says “Bantu”, without citing any Bantu cognates at all, much less a proto-Bantu form.

    The word is phonologically exceptional: Swahili doesn’t have closed syllables, and /l/ is never syllabic. If it is Bantu at all, it’s undergone some extremely odd sound changes. It may be borrowed from some other Bantu language, I suppose, but Meinhof’s handwaving don’t make it so.

  30. I was trying to find out whether the earliest attestations of words of the family of mfalme refer to non-Muslim rulers in primarily non-Muslim societies, and whether a Persian etymology was possible… On the way, I found:

    Nurse and Hinnebusch (1993) Swahili and Sabaki. A Linguistic History include mfalme in their Appendix 3 ‘Doubtful Proto-Sabaki Lexis’, p. 664 (class 1/2):

    ?*mufalume king (Seuta and PlO < Swahili): mufalume (Gi); mfaume (Am);
    mfalume ‘chief (Ti); mfalme (Mw,Ung); mfalume (Mn); mfa(l)ume (Ng)

    Google Books offers the following snippet from Documents comoriens II (1983), Institut national des langues et civilisations orientales :

    mfalme : origin unclear, distribution limited to Swahili, Comorian (mfaume) , Mijikenda (mufalume) , and a few smaller languages adjoining the coast in Tanzania. Since Comorian has borrowed a great many words from Swahili, but there is little evidence that vocabulary has regularly travelled in the other direction, Comorian mfaume may be an historical loan from Swahili.

    Mijikenda mufalume is glossed as ‘one who is violent, oppressive’. This suggests the possibility that the term has also been borrowed from Swahili, shape intact, but meaning modified to reflect a Mijikenda view of Swahili rulers.

    I was interestested by the following from Gill Shepherd (1982) ‘The Making of Swahili: A View from the Southern End of the East African Coast’, pp. 129–147 in Paideuma Bd. 28, From Zinj to Zanzibar: Studies in History, Trade and Society on the Eastern Coast of Africa:

    [Al-Idrisi’s account of 12th century East Africa] needs to be correlated, first, with what emerges from Comorian chronicles. Of these, I am aware of four. Two come from Nzwani, one by Said Ahmed Ali (1927) and one by Said Ali Amir (1961), neither published although duplicated versions exist, both of which have been quoted extensively by Robineau (1967). The two from Ngazija are both published, respectively, as The Swahili Chronicle of Ngazija by Said Bakari (ed. by Harries 1977), and Histoire de la Grande Comore by Abdul Latif Msa Fum de Mbeni (Tananarive: Imprimerie Nationale, 1917). The Ngazija chronicles both tell of a pagan period under rulers known as mafe (cognate with fumo/mfaume), followed by Muslim mabeja (Dubins 1972:22) — and Ngazija boasts a tenth century Muslim tomb (Viallard 1971: 177-78). The mabeja were in turn followed by Shirazis. The Nzwani chronicles tell of mabeja being followed by another group of rulers, fani, and only then by Shirazis. Both Nzwani chroniclers assume that the mabeja, as the earliest rulers, were pagan. I believe, however, that they were wrong. The title mabeja (singular beja) is derived from Persian, and had the connotation of “well-dressed person” (Johnson 1971: 267) — a noteworthy attribute in a world of costly trade-cloth. It seems unlikely that rulers who took their title from an Islamic language would have been pagans, especially if those using the same title on an adjacent island were Muslims. It remains to explain Nzwani’s fani. Their title, though said once to have been used by rulers in Ceylon (J. Martin 1968: 45), is unknown elsewhere in the Swahili world: it means “worthy” or “prosperous” in Arabic (Johnson 1971: 91). It is difficult not to identify them with Idrisi’s “Chinese”, and also with Ibn Said’s Southeast Asians who settled on Madagascar from South India at about this time, especially since the title beja survives among Nzwani’s uplanders (Hebert 1960: 103). The “Chinese” fani who drove the earlier (but still Muslim) mabeja into the hills were, we may conclude, the more immediate ancestors of Madagascar’s Zafi Ramni group.

    Is the assocation of fumo ‘chief’ and mfaume ‘king’ that Shepherd makes simply a Scheingleichung?

    Here are the two entries from Johnson’s Swahili dictionary that were referenced (asterisks indicate words on non-Bantu origin):

    Mbeja,* n. wa- a person who is neat, smart, well dressed, careful of ap¬ pearance. Mbeja wa kano, a fine muscular man, athlete. Ubeja, n. smartness, neatness, good athletic appearance in a man. (Pers. بجا [ba-jā] in place.)

    Fani,* a. worthy, fitting, prosperous. (Ar. فنع [faniʕ])

  31. David Eddyshaw says

    Mfa(l)ume at least looks more echt-Swahili, and there has been borrowing between dialects internally in Swahili, so that does at any rate diminish the problems with supposing that the word is of Bantu origin, at least.

  32. David Eddyshaw says

    The -fa- is peculiar, too. I’m not very clued up on historical Bantu phonology, but I think this can only have arisen (if it really is Bantu) by a tense high vowel having dropped out before the /a/, as with Swahili zaa “give birth”, from proto-Bantu *biada, and -fa “die”, cognate with the root of Kusaal kum “death.”

    A hypothetical *kuad- would actually be a reasonable match for the proto-Oti-Volta root *k͡pár-/k͡pér- that turns up in Kusaal kpɛɛm “elder, superior; official” and Moba kpéĺ “be older than”, but unfortunately looking for the etymology of a Bantu (?) word unique to Swahili by comparing it with Oti-Volta strikes me as somewhere between methodologically suspect and delusional …

    There is a proto-Bantu *-kʊ́dʊ́ “elder”, but the vowel quality is wrong, both for it to correspond to these Oti-Volta forms, and for it to give rise to Swahili f-. On the other hand, phonologically it is a good match for the root in Kusaal kʋdig “age, dry up, wither”, Mooré, kʋ̀ɩ, though the Oti-Volta forms here refer to the unpleasant physical side of this business of aging, rather than the delightful respect it so often engenders in the young.

  33. The Ngazija chronicles both tell of a pagan period under rulers known as mafe (cognate with fumo/mfaume)

    About this proposed cognacy… It seems that most scholars take Swahili -fumo (class 1/2) ‘chief; nobleman, knight; king, lord; bishop in chess’ as deriving from fumo (cl. 5/6) ‘spear with big flat blade, javelin, lance’. The latter has a good Proto-Bantu etymology, apparently—Guthrie’s comparative series no. 1867 -tú̧mò and no. 1868 -tú̧mù, both ‘spear’ (Comparative Bantu, volume 3 (1970), p. 136).

  34. David Eddyshaw says

    Surely the Swahili/Sabaki reflex of *tu- should be su-, not fu-? Or indeed, tu-. Still, as I say, I’m not well up on this, and the various Sabaki languages and dialects seem to have a bewildering variety of reflexes.

    Bastin’s proto-Bantu has *-túmo “spear” and *-kúmú “authority, kingship.”

    However, Nurse and Hinnebusch derive both proto-Sabaki *ifumo “spear” and *ifumo “chief” from *-fum- “stab, hunt”, but under the entry for the verb, which they do indeed attribute to proto-Bantu *-tum-, they hedge their bets by also citing *-kúmú. They describe the use of *ifumo for “chief” as a “traditional title”, which makes sense, as it’s Class 5/6, exactly like “spear” itself, rather than the “person” class 1/2.

    Simply asserting that fumo and mfalme are cognates looks pretty handwavy to me. I can’t see how you can get to *(mu)falume from *(i)fumo. It seems to boil down to “both used for ‘king’, both start with ‘f'”. Not convinced by the blithely asserted cognacy with mafe either. Given the multiple origins of Swahili “f”, it’s even weaker as evidence of cognacy than that.

    I notice that Swahili actually does have a (perfectly regular) reflex of the proto-Bantu *-kʊ́dʊ́ “elder” etymon: mkuu “chief.” So nothing to do with that, anyway,

  35. David Eddyshaw says

    No, I was wrong about Sabaki *tu-: nice table on p115 shows *tu -> fu as the regular development almost everywhere (it’s only *ti- that shifts to si.)

    Happily, this does not undermine my other gripes.

  36. Trond Engen says

    David E.: the various Sabaki languages and dialects seem to have a bewildering variety of reflexes.

    The Greek of Bantu?

  37. “respect it so often engenders in the young”

    Even in young people of today? You’re breaking my heart. I thought the world has already reached the moral bottom, but it still has a long way to go!

    (or maybe the stock image of elderly people complaining at lack of respect from young people of today is less common among English speakers than among Asia-adjacent Russians – аксакал “aqsaqal” entered Russian Central Asian vocabulary alongside with саксаул “saxaul“).

  38. David Eddyshaw says

    I gather that there are Egyptian texts from four thousand years ago, complaining about how the young don’t respect their elders any more.

  39. Even in young people of today?

    I believe DE’s habitual dry wit has escaped you. He was being sarcastic/ironic.

  40. David Marjanović says

    complaining about how the young don’t respect their elders any more

    And drawing the logical conclusion that the end of the world is imminent!

  41. David Eddyshaw says

    And how right they were! (I mean, what’s four thousand years or so in the overall scheme of things?)

  42. “The end is nigh” is simply wrong. It’s a misreading for The end is “neigh”.

  43. David Eddyshaw says

    A natural misunderstanding.

Speak Your Mind