By GJK Campbell-Dunn M.A. (NZ), M.A. (Camb.) Ph.D.

In this account we will systematically compare Basque morphology and sounds with Niger-Congo. The evidence will demonstrate that Basque, without doubt, is a Niger-Congo language. Basque is a member of the Togo Remnant (now Left Bank New Kwa), which was once a major Group. Basque is especially close to Akpafu.

The Togo R. people travelled by ship to the Canary Islands and up the coast of Spain to the Bay of Biscay where they established fishing settlements. The genetic data (Bertranpetit & Cavalli-Sforza 1991 : 55, Fig. 1, first principal component) locates Basque genetic material in a semicircular area around the Bay of Biscay, which was settled at least from the Palaeolithic Period.

The Niger-Congo cognates are taken mainly from Westermann (1927) and Mukarovsky (1977). For Basque our main sources are Trask (1997), The History of Basque, and Aulestia and White (1992), Basque-English English-Basque Dictionary. Other authorities are referred to in the text as relevant.


Typographical Conventions

On this web page schwa is represented by the character 3, and a palato-alveolar fricative by the combination zh.



Numerals are resistant to change over time and are a good indicator of language relationship. The Basque numerals contain some fossilised Niger-Congo morphology.

The word "one" bat in Basque, is related to Niger-Congo Nde i-bal, Mbofon e-ba, Guang ba "one", Gurma ba "one", Gola ba  and so on. (Johnston quotes forms with initial m in this word.) The final -t is a singular marker (things), as in Kebu (Westermann 1927 : 106): gobi-di "bone", e-ko-bir "bones" (with Basque -t from final -d). Fortis and lenis plosives are used grammatically in Kebu, a Togo R. language.

The word "two" bi, bi-ga, bi-da in Basque relates to Bantu bili "two" (Johnston 1919 - 22 : 32), and ultimately to Niger-Congo bi "breasts", which often has an l suffix. It is more closely connected semantically with Niger-Congo bali "two" (twenty-two reflexes in Westermann). The -ga, -da are "suffixes". Root da means "finger" in Niger-Congo.

Basque "three" hi-ru, hi-ru-r is similar to Etruscan ci "three", with k > h again (+suffix). This sound change h > #  is regular. Niger-Congo ki > hi is from kiti (Westermann 1927 : 288) "middle" (of the hand), which may be reduplicated to titi, kiki. Compare Avatime ku-ši "middle". The r- suffix is an old plural. Compare e-ko-bir "bones" cited above, with plural -r (< -t). For the change t > r in Togo R. see Westermann (1927 : 104). We take hirur from kitu-r. Compare Akpafu i-ru "five" with Avatime o-tu, Nyangbo i-ti, Kposso e-tu "five"   for the change of t to r. A phonological parallel with Akpafu in particular.

Basque "four" is lau, lau-r, Niger-Congo na "four". We have a sound change n > h > #  (with loss of what was once intervocalic n? after a prefix). An  l occurs in Nyangbo and Ahlo, both Togo remnant languages. Compare also Temne -aŋle (Wilson), Igala ele, Ishekiri meere (Armstrong 1967 : 63). The base form is na. The same plural suffix r may occur as in "three".

Basque for "five" is ba-st, bo-r-tz. This is Niger-Congo bua (> bo) "hand", and identifies "five" with the hand, since the hand has five fingers. Niger-Congo extensions of this word occur Efik bok, bak, Mossi bonko, baγado, boγo, boγo-re (with old plural suffix) Gurma boγu, Kissi boka. Bantu has voko "arm". The final -t (singular) is due to the "hand" being singular, though having five fingers.

The word for "six" is sei-r, sei in Basque (as in Italian). The form seir has the old plural in -r.  "Bantu" (Johnston) has sambω, sasatu, sasaba. Niger-Congo ta "hand" should be considered as the source. This root goes to sa in Tschi, Agni and is assibilated in Baga to ke-tsa. Delo has n-sia.

Basque zaz-pi is "seven", Johnston's "Bantu" sambω, sambω-bali. Similar considerations apply. The za- is for ta, sa "hand". The -pi is bi  "two". See Westermann (1927 : 110, 9), bi (> pi). "Seven" is 5 (hand) + 2.

Basque "ten" ha-ma-r is from Niger-Congo ka "hand" + ma "end", and refers to the end of both hands [-r plural ].

"Eleven" is ha-me-ka, "ten big", hamar + ga "big". "Twelve" is ha-ma-bi, "ten + two".

"Zero" or "nothing" is Basque hu-ts, from Niger-Congo ku "thing", compare French rien "nothing", from Latin rem "thing". The same word is used for "one" in Etruscan.

The Basque for 1000 e-hu-n is allegedly from Gothic.  The prefix e- often precedes this word. Root ku means "big" in Niger-Congo, then "big number", i.e. "ten" or "one hundred". Bowili kuwa "ten", Ewe kpi "much", Igbo okpi "big", Bulom akpil "many". Niger-Congo i > Basque u. Bantu kumi "ten". A Niger-Congo solution is to be preferred.

The attachment of suffixes to the numerals is a Niger-Congo feature: bat, batzu (pl), bi-ga "two" (ga "big"), hiru-r "three", lau-r "four", bost, bort-z "five", sei-r "six". The attached -r of hirur, laur etc. is a Niger-Congo plural (from ri), as in Etruscan. This is clear from the fact that the r is added whenever the plural suffix -ak follows: hirurak, laurak. Compare the names Aquitani, Liguri.

Niger-Congo numerals were once prefixed also, which tends to create phonological problems of identification. This applies e.g. to lau "four", often nna in Niger-Congo with an assimilated prefix, which has changed n to l. (Westermann 1927 : 104, n > l in Togo R.) and to bost "five", where the Niger-Congo once had a preceding labiovelar as in Guang o-ma-po "end".

The ordinals have a suffix -ren, len < Niger-Congo na  "be", "progressive" (Etruscan ordinals have -na "to be", "progressive"), e.g. biga-ren "second". But "first" is len( en) go, with suffix -ko < -go  (< Niger-Congo gwa "much"). The re, le is an old Niger-Congo article, often le in Togo R.

Entwhistle (1962 : 25) points out the vigesimal arrangement in Basque using hogei  (< kua "foot" + gi "two") as base. Vigesimal numerals are a feature of Niger-Congo (Gur, Yoruba), due to counting on hands and feet (before shoes were in use). Celtic took its vigesimal numerals from the same source. Latin adopted vigesimal numerals from Etruscan (XX, XXIV etc.), which we have shown by comparative linguistics to be Niger-Congo.


Personal Pronouns

Personal Pronouns tend to be well preserved over long periods of time. Basque for "I" is ni. The word ni "I" is omnipresent in Niger-Congo  and occurs across six language groups (fourteen reflexes in Westermann 1927 : 264 - 265), including Bantu. Basque ene   "my" (Trask 1997 : 246) has an ancient prefix and case inflection.

The word for "I" however is also ni in Hausa (Chadic), which is spatially close to Berber. Berber for "I" is n(3)ki, which may be from ni-ki, as it will be remembered that in Berber short i becomes shwa or zero.

The conclusive detail here is the existence of forms of the pronoun enlarged with suffixes: ni, nik, niri (eni), nire (ene), nitaz etc. Niger-Congo ni "this" has a form preserved in Dagomba nine. Compare for this suffix Bamana nine "earth" also from a word ni. One can add the evidence of Niger-Congo ni "name", Ewe niko, Grebo nine, Efik enin (with prefix). The dative and genitive of this word in Basque, eni and ene respectively (Trask 1997 : 97), appear to have a Niger-Congo prefix.

The affix gan"of local cases may be attached directly to the absolutive form" (Hualde & Urbina 2003 : 179) of Basque pronouns: nigan, niregan etc. Compare the Niger-Congo postposed ga, Ewe gama, Guang  ganom, ganen (Westermann 1927 : 213) meaning "here, there, to, towards, place, this place". This is a very significant detail.

Basque for "you" (intimate sg) is hi, which we relate to Bantu ku "you" (Johnston 1919 - 22 : 32), Niger-Congo (Westermann 1927 : 235) ku, Ewe ku, Ga, Grebo, Igbo, Likpe, Santrokofi, Animere, Kpossi, Bamana, Boko, Kpelle ku  "company, with, and, to help, to meet, companionship". Tobote has the form kute (compare Basque hitaz). The change u > i is common in Niger-Congo (and Basque).We derive Basque h from an earlier k. The second person "you" is a companion.

Hayward in Heine & Nurse eds. (2000 : 88) quotes forms of ni "I" from Afroasiatic, but reconstruct i/yi as the Afro-Asiatic protoform. For "you" (sg) he gives ki as a possible Proto-Afro-Asiatic form.

The Basque second person plural "you" (pl) zu, zu-ek, with plural suffix,  derives from Niger-Congo gu, i.e. *ghw u "he", Bantu γyu, γu , with the old notation for the aspirated (lenis) labiovelar, which has been assibilated in Basque (as in Etruscan). Westermann (1927 : 224 - 5) has thirty-two reflexes for this word, mostly wo, u, o.  The third person is often used as a formal second, as e.g. in Spanish and Italian.  Johnston (1919 - 22 : 32) gives practically the same *"Bantu" forms for "Thou" and for "He (she)". It should be added that Niger-Congo originally did not have different forms for singular and plural. See Entwhistle (1962 : 25) regarding "degrees of distance from the speaker".

Hayward gives ina, na, nu, ni as Afro-Asiatic for "we" however, which are not related to Basque gu (but to Basque ni "I"). His "you" (pl) is reconstructed as  *kuuna. Cf. Basque hi > ku "you" (sg).

Intensive pronouns (Trask 1997 : 97) add haur, e.g. nihaur . I suggest this formation is from Niger-Congo ka "also, and", with k > h, + plural -r.

Basque lacks third person forms "he, she" of the pronoun. But an early *d or *da (whence the pronominal prefix d-), with a variant l, also *b, *be have been reconstructed (bere "his", i.e. bere "of him"). The *b is from Niger-Congo ba "somebody, someone, one" (Westermann 1927 : 202, with eight reflexes). The *d is probably the Niger-Congo article and pronoun la meaning "this, he, she" (Westermann 1927 : 247, with seventeen reflexes). None of Westermann's forms has a d . But voiced initial consonants are a feature of Basque, though initial d is rare. For the 3rd person pronoun of Proto-Afro-Asiatic Hayward gives forms beginning with *s (which might come from a dental).

Genitives of the pronouns are: nere/ene, hire, zure, gure, zuen , all possessives (Lafon 1960 : 78).


Definite Article

The same word is probably behind the Basque article in postposed -a. This is reduced to -a in some Niger-Congo languages, pace Trask (1995 : 88), who says it "cannot safely be projected back to the pre-Roman period". The Roman period suffices for me. Kongo (Stapleton 1903 : 50, par. 161) has an article in a, o, e, almost certainly related to the West Sudanic article in la, ala, (na). The Niger-Congo "article" has a wide semantic range "the, this, that, he, she, possessor, 'bildet nomen agentis' " (Westermann 1927 : 247, with seventeen reflexes). The same is true of Basque -a  (= earlier  distal demonstrative -har, says Trask 1995 : 88), "whose use is much broader than the name would suggest" (Trask 1997 : 89). This is from the form na, with intervocalic n > # in Basque.. Sudanic languages also reduce this "article" at times to -a.


The Interrogative

Trask (1997 : 98) lists interrogatives nor "who?", zer "what?", noiz "when?", non, nun "where?", zelan "how?" The Basque forms with n- are related to Kongo nani?, Bangi and Lolo na, Ngala njai?, Poto, Ngombe, Sokonda?, Kele ndai?, Swahili nani?  "who?" (Stapleton 1903 : 102), and Western Sudanic la, na, no "this, that" (Armstrong 1967 : 65), Dagomba no "which", Kussassi o-no "that", Likpe nua "which", Akpafu ngo "which". The *a > o, u. The Basque forms with z- ( a voiceless sibilant) indicate an original lenis labiovelar as seen in Swahili kwa, Bangi kwe (Stapleton 1903 : 106) from Niger-Congo kua/khwa "man, person" (Westermann 1927 : 240). Compare Etruscan s > Niger-Congo *GHW and *KH. The Etruscan letter X (< Niger-Congo *KH) was a sibilant.



The relative (Trask 1997 : 114f) is added -(e)n, which is Niger-Congo -na, -ne "this" (thirty-three reflexes in Westermann 1927 : 261) and used as a relative in Bantu. Also -(e)la, which is the Niger-Congo article/pronoun.



The Basque conjunctive eta, ta "and" is Niger-Congo ta "dass, damit" (Avatime, Nyangbo), more often meaning "speak, tell"; Igbo ita "proverb". The exact meaning of this word in Basque is found in Ngala nta "and", which is also used as a superlative. The word ala edo  "or" is the Niger-Congo article la + prefix, which also means "this" used with the word for "one", do, likewise prefixed.



The prefixed negative ez- is Niger-Congo na, ne "this", "not", with loss of initial n and an added suffix. Newole, Grebo, Igbo na "not", Edo ne "not", Ahlo, Agni, Kposso na "not", Dagarti na "not yet", Soninke na "not". This word loses the n also in Etruscan. [Latin ne may be the same word.]  The z may be a vestige of ta, te "not".



The Basque word for their language is Euskara or Euskera, earlier Heuskara. The first part of this word is the Togo R. word for "Akpafu", Likpe be-fu "Akpafu",   Bowili o-vu-ne "Akpafumann",  Santrokofi o-fu "Akpafumann", Akpafu ka-wu, ka-'u "Akpafu". The early initial Basque h is from k, as can be seen from ka-wu, ka'u. The a has changed to e in this lexeme. The consonant between e and u has been lost. Basque lacks the semivowel w, which drops out here in Akpafu ka'u. See Lafon (1960 : 92) for confirmation from placenames etc.: Ausci, Aoiz, Auch.

The second part of the word, ka or ke is a word for "speak", Niger-Congo gue "voice, language", Ewe, Ga gbe "voice", Agni guere "language, speech", Yoruba i-gbe "loud cry", Gbari e-gwe, e-gbe "mouth". The e is for original a in this word. Niger-Congo e is secondary. Compare Niger-Congo ka, ke, k'e "to speak", which is related. The final sylable -ra is the Niger-Congo article. No clearer proof could be found that the Basques were originally the Akpafu! The parallel word Euskadi "The Basque State" is "Basque language" with the Niger-Congo word di "root, tribe" attached.

Both Basque (Latin Vascones) and Bastulan begin with the well-known Niger-Congo prefix ba-, used to form names of Peoples. Lafon (1960 : 92) cites a form Barscunes from coins. Compare Bantu, Bassa, Bamana, and especially Bargu = Barscu-nes. For Togo R. see Westermann (1927 : 110, 9), ba- "Mz von Menschen".



In Niger-Congo numerals, pronouns and nouns share the same prefixes and suffixes. This is true also of Basque, but in Basque the prefixes are largely vestigial, as in Etruscan. 

Many Basque words begin with a vowel, which can be shown to be a Niger-Congo vocalic prefix. Compare Martinet (1964 : 379) regarding the abnormal frequency of initial vowels in Basque. Martinet  (1964 : 383) suggests an old Basque stress accent on the second syllable of the word. Prefixes tend to be unaccented. Noun prefixes are the hallmark of Niger-Congo languages, and are crucial to our proof that Basque is Niger-Congo. Western Niger-Congo had a stress on the root (second syllable), but also a pitch accent.


Noun Prefixes

The prefixes of Togo R. are listed in Westermann, with their functions, many of which have merged. Examples (see especially Akpafu) in Westermann (1927 : 112 - 121): a, e, i, o, u, le, di, bo, bu, ba, be, bi, k, ti, s, f.  Judging from Bantu all Niger-Congo prefixes once began with a consonant. We organise the nouns by prefix, but cite suffices also where possible. Togo R. suffixes include (Westermann 1927 : 108 - 109): e, u, a, li, ri, r, ni, mi, bi, k, t, si, s, f.

The meaning of these elements is obscure, but they may have originally matched and copied the prefixes, as in Gur. The categories of Living Creatures and Things are common to both. Some of these prefixes and suffixes undergo sound change or loss in Basque, which allows a very limited range of initial and final consonants (Trask 1997 : 88). Westermann puts Togo R. close to Gur (and some Kwa languages).


Prefix a-

Basque ama "mother" is well established for Niger-Congo: Indiki, Korop, Kebu, Konyagi ama, which all have the root ma (twenty-nine reflexes in Niger-Congo according to Westermann 1927 : 254) with the same prefix. Some however object to the use of "mother" in comparisons on the grounds that the available words for "mother" are limited by the vocal organs of small children. Trask (1997 : 269) "probably of nursery origin". I regard multiple cognates as probative, and terms of family relationship as valid evidence. This is particularly true when several connected words are involved and one can identify morphology. The Basque word(s) for "father" can also be explained as Niger-Congo.

Basque aita "father", Niger-Congo ta "father", Ewe ata, Igara ata, Okpoto ata, Edo ita, Nki, Uwet, Djuku ata, Nde, Animere nta, Djola ata "chief, king", Bantu -ta "father".  Reduplicated forms tata in Nso, Kpossi, Malinke; Bago tatana. The a-prefix sometimes has a following -i- in Niger-Congo. Entwhistle (1962 : 18) takes aita as "possibly" from Celtic (Oir (p)athir, which can be rejected on phonological grounds. One might as well cite Latin pater. Compare Aquitanian Atta-, which "possibly represents this word" (Trask 1997 : 269).

Trask describes the other Basque word for "father", aba, as an "eccentric 19th century neologism" (Trask 1995 : 89). But it has a good Niger-Congo etymology in ba "father", Dewoi, Gbe ba, Yoruba oba "king, father, sire", Kukuruku oba "king" [with a- > o-], Gurma oba "father", Banyun aba "father", Gan (Mande) aba "father", Bantu -vava "father". In fact twenty-five reflexes in all.

Basque alaba "daughter" appears to have the same prefix, as does ahizpa "sister": (of a woman) < *anizba. This is from Niger-Congo na "mother", with n > l (an Akpafu change).  The suffix ba, pa occurs elsewhere. "Sister" (of a man)  arreba follows the pattern, having the same prefix and suffix.

To these we can add anaia "brother" (of a man), which Trask derives from a word with a fortis (double) nasal N. We have the same prefix. The word is from Niger-Congo nia, nua "brother-in-law, relative, kinsman" (Westermann 1911 : 169), Ewe no, Tschi oni, Yoruba ana. Westermann gives a palatal nasal for Ewe here. The word ahaide "relative" is from the same root (< *ana + ide "fellow", according to Trask 1997 : 269). Probably the words for "daughter" and "brother" are connected with words for "mouth", Niger-Congo da, la and nua, nia (kissing).

Basque agure "old man" has the same prefix again and is from Niger-Congo gu, gua, ghwa "grey hair", Ewe wo, Ga wan, Yoruba ewu, Igbo awo "white hair", Efik iwat. There is a problem here as Niger-Congo gw > Basque  k and Niger-Congo ghw > Basque s. Trask quotes Latin AVULE. "Old woman" atso may be from Niger-Congo tua > to > so "water", Guang ntsu, from the habitual practice of fetching water assigned to (old) women.

Basque asaba "ancestor" is from Niger-Congo GHIA "blood", gia, Ewe kadze, Tschi boga, Guang obudza, Nupe egia, Gbe aga, agya, Lefana ubudza, Ahlo obidza. Basque s > Niger-Congo GHI. The suffix -ba may be "father".

The word for "goat", Basque aker, has the same prefix a-. But the k is from a Niger-Congo labiovelar gu, gw seen in Ewe gbo, Dahome u-gwa (different prefix), Ani, Baule, Afema bwa, Ga gba-ten, Ake gba, Sya gwa, gba, Tschi o-gwan, Mekibo ebwane, all of which mean "sheep" or "goat". Basque *k > h > #. Basque *gw > k.

The root a of the word has changed to e in Basque. Compare Etruscan a > e.

Basque astun "grave", Togo R. languages Akpafu karo "grave" (with s > r ), Avatime otsro, odro, odzro "grave", Guang ntsa "grave", all from Niger-Congo li, di "to dig", Ewe di "to dig", Efik udi "grave", Likpe kudi "grave". Basque has t for d and u for i here.

Basque adar "horn", Togo R. languages Logba esa "horn", Avatime lisiane "horn", Guang esiebi "horn", from Niger-Congo ta "war", horns being associated with war.

Intervocalic stops tend to be voiced in Basque. Thus t > d.

Basque adar "branch", Niger-Congo Togo R. languages Logba ula "branch", Likpe ole "branch", Santrokofi ola "branch, bough", Ewe alo "branch". Basque has d for Westermann's l here, a common development.

An example with prefix a-  from Vidos (1959: 221) Basque ara-otz "cold plain", Niger-Congo la "earth"; [Basque r = Niger-Congo l], Ewe de "on the ground", Igbo ala "earth", Yoruba ile "ground, earth", Bulom alo "under, below", Santrokofi kala "the below" (this last retains the original initial consonant in the prefix). [For -otz compare the placename Ara-güés.] Basque landa "plain, prairie" is from the same root la "earth".

From Vidos (1959: 232-233) we can cite Basque agogai "opening" (compare Pliny NH XXXIII, 69-75, who mentions agogae as meaning "opening in a gold-mine"), Niger-Congo gua, gwa "gate, opening", Ewe agbo "door, entrance to a house", Ga agbo "gate", Guang gegba "tower", Yoruba abgo "the inner court of a house", Gbari gwami "mouth, opening", Edo ogbolo "forecourt of a house", Boko gba "door".

The Basque word arroil "ditch, narrow mountain passage", arrugia, "drainage ditch, water conduit, canal"  (compare arrugia "canal, passage in a gold-mine" quoted by Pliny Ibid) is related to Niger-Congo gia, gwia "water", Yoruba ogi "water", Igbo igi "water, spring", Akwa aya "river", Bowili koya "river", Adjuku midji "water", Malinke dji "water", Mende yia "water", etc.

Basque azal "skin" (Trask) is from Niger-Congo ge, gel, *ghwia "skin", Ewe γe, we, wo, Dahome wema, woma, Tschi were, nwoma, Ga wolo, Guang were, owelo, owolo "skin", Beri kawolo "skin", Abure awe "skin", Kya ahoro "skin", Igbo iwolo "cast skin", Bowili owori "skin". The lenis voiced labiovelar of Niger-Congo becomes Basque z [predorsal fricative]. Forms with o due to labial influence.

Basque arrain, arraki, etc., "fish" is from Niger-Congo ta "to catch", also "to trade" (trade in dried fish), with the Togo R. (Akpafu) change of t > r (Westermann 1927 : 116, no. 77). Ga ta, Guang ta "to catch", Gurma ta "to catch", Senufo ta "to have, to find", Bamana, Djula, Soninke ta "to catch". The lack of a Proto-Niger-Congo word for "fish" (PWN KHIUNI, KHUINI, Bantu cui [Meeussen]) may be due to the use of the verb "catch" for this concept. Compare Bantu dob "catch fish", dub "fish with net".

Basque amuzki "fly", Niger-Congo (m)bu "bug, mosquito", Ewe mu "mosquito", Bassa mui "mosquito", Yoruba ure "mosquito", Igara imu "mosquito", Ebe amu "mosquito", Egbo anu "mosquito", Eka abu "mosquito", Djuku mu "mosquito", Boritsu emun "mosquito", Afudu amua "mosquito", Bantu -vu "mosquito".

The prefix m changed bu to mu and a new prefix was added. Akpafu mb > m.

Basque amai "end", Niger-Congo ma, man "to be at an end", Ewe mamle "remains", Guang omapo "the last", Efik ma "to finish", Ekoi man "to finish", Avatime moni "to end", Djan ma "to be finished", Bulom mani, man "to leave off", Bantu mala "to be at an end".

Basque amu "fishing net", Niger-Congo mu "to catch", Mekabo mu "to catch", Yoruba mu "to catch", Okpoto mu "to catch", Oloma mu "to seize", Bowili mu "to catch", Akpafu moe "to catch", Santrokofi mu fu "to seize".

Basque aho "mouth", Niger-Congo ka "speak", with the noun from the verb. Compare ka "bite".


Prefix i- (common in Akpafu)

In western Niger-Congo these noun prefixes are omissable. Hence Basque bai, ibai, ibaiko "river bank" (Vidos 1959 : 221). This is from Niger-Congo ba "mud, slime, clay" seen in Ewe ba "mud", Edo iba "mud seat", Efik mbat "clay, mud", Logba oba, Dagomba ba mud, pond", Mossi baka "sea", Temne abat "swamp for rice planting", Gola eba "mud", Mende bati "swamp".

Basque iturri "spring", Ewe to "water', Tschi nsu "water", Afema asue "water", Agni nzue "water", Zema azure "water", Guang ntśu' "water", Ake su "water", Likpe mtu "water", Lefana ntu "water", ketu "river". A frequent change of  to s in this word.

Basque idi "ox", Ewe lu "cow antelope", Abe lu "cow", Avatime olu "antelope", Ahlo ulu "cow antelope", Gola odi, oli "cow", Mano di "cow", Gio du "cow".

Basque izar "star-shaped mark on the forehead of an animal", izardun "animal which has a star on its forehead", compare Iberian *izar-di "animal with a star, chamois", Niger-Congo guani "sheep", Dahome ugwa "sheep", Tschi ogwan "goat", Avatime bwa "sheep", Zema bwane "sheep", Afema bwa "sheep", Ga gbanten "goat", Ada agbani "sheep", Mekabo ebwane "sheep". The chamois resembles a goat. Niger-Congo final n changes readily to r.  Note that Berber has išri "star". Niger-Congo aspirated (lenis) labiovelar gives Basque z (a sibilant), as in Etruscan.

Basque ile "hair", G. dial. Ule Niger-Congo li "head", Yoruba ori, oli, Okpoto adzhi, Igbo elu "top, summit", Akparabon eri "head", Lefana kuli, Ahlo ilo, Bowili lelo, Kpossi elo.

Basque ipurdi "buttocks", Niger-Congo bua, buo "hole", Ewe vo "hole", Efik ibuo, ubuo "nose", Igbo ubwo, ubo, ugbo "hole", Gbe obwo "hole", Ibibio igwo "nose", Delo bo "hole", Kussassi boko "hole".

Basque izerdi "sweat", Niger-Congo gia, ghia "water", Igbo igi "water, spring", Mbofia igi, idzh "water", Adjukru midji "water", Akwa aya "river", Bowili koya "river", Mende yia, Sussu ye "water".

Basque izter, aztal "thigh", Niger-Congo ta "thigh", Yoruba ita, ito "thigh", Yala itaku "thigh", Kamuku uta "thigh", Igara ota, Eafeng ota, Ahlo ote. [it- > ist-.]

Basque ipurzulo "anus", Niger-Congo Togo R. languages, Nyangbo kipi "anus", Adele opi "anus", Avatime opino "tail", Logba ovi "tail", Ani bi "dung", Nki abun, Ekoi abin, abun "excrement", Ahlo ebi "dung", Grussi bun "dung" (dial.), Bantu vi "excrement". Basque has u for Niger-Congo i.


Prefix e-

Basque etxe, etxa- "house", Likpe le-tsya, Bowili le-tsa, Akpafu i-sa "flat roof", Logba u-tsa "house", Ewe ta "top, head", Yoruba a-ta "the ridge of a house", Guang ta "to dwell", Animere ta:to dwell, to remain", Ga ta "to sit". The Basque palatal fricative x is from an earlier sibilant. The root vowel a is earlier than e. This is one of the essential pieces of evidence that Basque is from Togo Remnant.

Basque eki "day", Niger-Congo gui "sun", Tschi ewi "sun", Guang owi "sun" Mekabo eyileyi "sun", Bassa giro "sun", Nupe eyi "sun", Akwa nyui "sun", Nyangbo kiwi "sun", "day", Logba adzhi "day", Likpe diyi "day", Akpafu iyi "day", Adele duwi "day", Santrokofi diyi "day", Ahlo iyi "sun", Kposso eyi "day", Konkomba nwi "sun", Gola egwe "sun", Ewe (Clarke) dji "sun". Basque lacks semivowels (except j- ?).

Basque esku "hand", from gua, gwa "hand, arm", Nupe egwa "arm", Yoruba owo "arm", Igara owo "arm", Okpoto owo "hand", Bowili awoe "hand", Kpossi uwo "arm".

Basque euri "rain", Niger-Congo tu, tua "water", Tschi nsu, Afema asue, Baule nzue, Zema azure, Guang ntsu, Beri ntjum, Ake su, Gwa ndu, Kyama ndu, Lefana ntu, Akpafu ndu. Bantu -dumb (Meeussen). We assume loss of intervocalic t (t > tr > #) in this word. Forms with d are due to a preceding nasal.

Basque eultzi "faggot, bundle" but also "thresher". Niger-Congo tu "to gather, put in heaps", Yoruba tu "collect, gather together", Gbari tu do "to heap up", Igbo tu ba "to heap togther", Efik ottu "of things heaped together", Malinke tun "put in a heap", Bantu tunda "heap up". Same sound change as previous example. The meaning "thresh(er)" is from Niger-Congo tu "knock, strike, push, pound rice" etc. Temne tun "to knock, strike, push".


Prefix -u

Basque ura "water", Niger-Congo gia, ghwia "water", Yoruba ogi "water" (dial.), Igbo udzhi "water", Akwa aya "river", Bowili koya "river", Kpelle ya "water", Mende yia "water", Sussu ye "water". The prefix induces a change ghw > r, also seen in Linear A (RA "snake" < GHWYA).

Basque ume "child", Niger-Congo bi "child, Ewe vi "child", Ga ba "child", Guang obi "child", Gwa bi "small", mi "child", Kyama mi "child", Gbari ebi "child", Mussu bi "child", Edo obi "child", Avatime obi "child", Konkomba mbim "child". Trask notes that the Basque m comes from a b here, which is highly significant. Westermann (1927 : 104, 174) gives examples of b > m in Togo R., including mi > bi "child", Kposso u-mi < m-bi "faeces". Bantu has viala "to bear". Westermann (1927:  207) has fifty-eight reflexes for this word, which is regarded as an early lexeme. Compare muan "to bear child", oma "child", etc.

Basque uda "summer", Niger-Congo da "day", Tschi eda "day", Dagarti da "day", but also Ewe da "always", Tschi da "always", etc. Compare udamin "dog days, hottest days of summer", with Niger-Congo min "to drink". Also udaberri "spring" with addition of berri "new". Compare Egyptian ra "sun god" [r = l ].

Basque upa "a vat", Niger-Congo ba "a yam store", Yoruba aba, Igbo oba.. Also Ewe ava "storehouse, barn". Compare Niger-Congo pa (I) "full", Dagarti pa-le "to be full", Boko pa "to fill, be full".

Basque uraldi "flood", Niger-Congo gi, gia, ghia  "water", Yoruba ogi, Igbo udzhi, Akwa aya "river", Bowili koya "river". Basque urtegi "tank of water", from Niger-Congo gi "water". The base word is ur, ura "water".

Basque urtoki "swamp", "water place", Niger-Congo tua, to "water", Ewe to, Tschi nsu, nsuo, Afema asue, Agni nzue, Guang ntsu, Obutu ntšu. Compare urtsu "watery", urtu "to melt", with ur "water" + tsu "water".

Basque usadio "custom", usaia "usage" Niger-Congo ta > sa "to be accustomed", Guang ta "to be accustomed", Ga ta "to sit", Ahlo ta "bildet Habitual" W., Animere ta "to remain".

Basque urkabe "scaffold, gallows", Niger-Congo gwa "scaffold", Ewe agba, Ga agba "a small scaffolding", Grebo gbabe "a bier", Temne agbaya "shed" (flat roof). Related urkari "hangman", but also "water carrier" < gwia "water".


Prefix o -

Basque ordu "hour", Niger-Congo tu "hour", tsi "to hear" (of hearing a clock strike), Togo R. languages: Avatime kutu "hour", Nyangbo ato "hour", Likpe koto "hour", Ahlo oto "hour", Lefana kotu "hour", Kposso utu "hour".

Basque osaba "uncle", Niger-Congo Togo R. languages, Bowili ofa "uncle", Akpafu ofa "uncle", Santrokofi ofa "uncle", with Togo R. f = Basque s. This s is from earlier t, Niger-Congo ta "father" with Niger-Congo ba "father" added. Uncles were classified as "fathers" in parts of Africa.

Basque ondo, onto "fungus", Togo R. languages Akpafu orui "fungus" (s > r), Likpe usu "fungus", Guang itsuli "fungus" < *ii-tsuni < tuni. Basque has interchange of d and t here, and change of u to o.


Consonantal Prefixes

Basque zitu "grain", Tschi, Nupe eti "head", Konguang nti "head", Djar mo-ti  "head", Likpe, Santrokofi disi "head", Akpafu iti "head", Bantu tu "head". The word is used for a head of grain in Western Niger-Congo languages, and also in Linear A. Sound change in Basque i > u. Basque z is a sibilant.

Basque sutondo "hearth", Niger-Congo ta "fire", Grebo to "fire", Edo eta "fire", Santrokofi oto "fire", Akpafu oto "fire", Likpe oto "home, hearth", Bowili oto "home, hearth", Adele otan "fire", with a > o.

Basque zidor "road", Togo R. languages: Akpafu ori "road" (s > r), Likpe kusu "road", Bowili osi "road", Avatime kuto "road", Niger-Congo tu "to go away", Tschi tu "to emigrate" etc., with Basque t > d. Basque z is a sibilant. This sibilant comes from an original t. Basque changes u to i.

Basque buru "head", Grebo lu "head", Gbe duru "head", Yoruba ori "head", Igbo elu "top, summit", Lefana kuli "head", Kpossi elu "head". Kasanga (Atlantic) has bugof  "head" (Wilson in Bendor-Samuel ed. 1989 : 99), with the same prefix as Basque. Nyamnyam (N. Bantoid : Hedinger in Bendor-Samuel 1989 : 424) has biri "head", allegedly an innovation. For prefixes with b- in Togo R. see Westermann (1927 : 110, 8, 9). Basque changes i > u, so the prefix is Togo R. bi.

The syllabic prefix is retained in Basque, also in Gbe and Lefana. In view of the k in Lefana an original labiovelar is probable. In early Basque lenis plosives were reinterpreted as voiced (Trask 1997: 166). We posit a pre-Basque lenis labiovelar GHW as the source of *b-. This is responsible for the initial o- in Yoruba.

[Basque mutil "boy" (< Latin PUTILLU: so Trask 1997 : 271). The word may ultimately be from Niger-Congo ti "head", Djarawa mot "head", or Niger-Congo li > di "head", Yoruba oli "head", Lefana kuli "head", with the Basque devoicing of the fortis stop, or from Niger-Congo li "root", Guang oli "root", Animere kuru "root", or Niger-Congo ti "tree", Bowili kutsi "tree". Umbilical cords were buried under trees in Africa when a baby was born. This assumes an original labiovelar prefix, as in Swahili mu-toto "child" < *gwu-. There was Niger-Congo influence on Latin, which came via Etruscan. Tene (Koelle) has didi "boy", which supports original d here.]

Basque bekoki "forehead", Niger-Congo kua "neck", PWN KWANT "neck, nape of neck", Ewe ko "neck", Tschi ekon "neck", Yala oko "neck", Likpe okwe "neck", Ahlo ogo "neck", Kono ekana "neck", Toma ko "neck", Sussu koni "neck".

Tschi and Kono have a vestige of the syllabic prefix with initial lenis labiovelar that has been retained in Basque as be-. This is responsible for the k- of  Sussu and for the rounded vowel of the prefix in Likpe and Ahlo.

The forehead and the neck were both used for carrying burdens in Africa. A strap could be worn around the forehead for this purpose.

Westermann (1927 : 110, 187, 9) notes that in Togo R. the ba, be, be prefixes are used for plurals of people, animals, things. The prefix ka is used for diminutives, he says, in Akpafu (as in Bantu). Notice that Basque hatz is "finger", but there is also a form behatz "finger". For "toe" we have only behatz, however. "Eye" is begi. For "hair" we have ile and bilo. For "hand" we have esku and bosteko. For "leg" hanka and zango. For "knee" however only belaun. These Basque forms in ka are old diminutives, whereas the forms with prefix b appear to be old plurals which have become fossilised singulars. The plural suffix -r becomes likewise fossilised.

On page 106 Westermann indicates that Kebu formed the plurals of nouns for persons and things by voicing the relevant initial consonant. Likewise Avatime-Nyangbo, Logba, Lefana, Bowili, Ahlo, Kposso. This may be pertinent here. See Westermann's comments on page 107 for examples and an explanation. In general Akpafu does not have this feature as part of the system, but it may be very ancient.

Basque burduña, burdina "iron", Niger-Congo, Kono, Vei kundu, Soso wure, Dewoi wulie "iron" (Koelle), Idsesa ure "iron" (Koelle). Again with labiovelar prefix. Compare tu "iron", and Bowili dina "extinguish".

Basque labaina "knife", Niger-Congo Togo R. languages, Bowili lepame "knife", kopa "bush knife", Avatime kapami "bush knife". Basque has b for p in medial position (originally a Basque fortis consonant medially).

Basque ganadu "cow", Niger-Congo na "cow", Guang kena "cow", Obutu ona "cow", Edo ena "cow", Koro ona "cow", Djuku na "cow". Kamuku bina "cow", Tara na "cow", Padjade kuna "cow". The ga- is a prefix, the -du is from lu, du "cow". Westermann (1927 : 262-263) has fifty-three reflexes for this word. We consider that the cow was a very important animal for speakers of Niger-Congo.


Without Prefix

Western Niger-Congo languages are remarkable in that prefixes on the noun appear optional. Forms with and without a prefix coexist. Some examples without prefix:

Basque zur "wood", Niger-Congo tu "tree" > su, Grebo tu "tree", Newole su "tree", Efik etu "tree", Kpossi itšu "tree", Foro tiri "tree". [t > s]. Kru su "tree" (Marchese in Bendor-Samuel 1989 : 131). Basque z is a sibilant. Added suffix -r (Westermann 1927 : 108, 183, 184, 3).

Basque oin "foot", Niger-Congo kua "to go", Ewe ko "to go", Ani ko "to go", Mekabo ko "to go", Igbo ko "going", Efik ko "to go", Soninke kokoi "to go", Kissi kon "to go", Gola ka, ko, kwa "to go", Bantu ka "to go"; k > h > # suffix -n (Westermann 1927 : 108, 184, 1).

Compare Niger-Congo kua "leg, foot", Newole kpole "foot", Efik ukpot "foot", Lefana ogba "leg", Numu kpo "foot", Huela kpo "foot", Gagu kpa "foot".

The notions "foot" and "go" are related. The same symbol is used for both in hieroglyphics for example. The kp palatalised stop is lost in Basque.

Basque haragi "flesh", Niger-Congo ka "flesh", Guang okam "fish", Nyangbo aga "beast", Ahlo ika "beast, flesh", Bowili sekana "flesh", ka "beast", Bamana kara "flesh", Foro kari "flesh", Takponin kara "flesh", Temne oka "beast". The ending -gi of this word may be from gi "blood", or gu "kill, die".

Basque hilargi "moon", Niger-Congo kia "to become bright", Ewe ke "to become day", Tschi k'e "to become visible", Animere etje "day" + la, da "day", [+ argi "light" (Trask 1997 : 312)]. The word is used of the moon and stars.

Basque mintz "skin", Niger-Congo Togo R. language, Nyangbo bupi "skin", Ewe fi "thin skin", with the Basque sound change of b, p > m.  The same change occurs in Basque ume "child", < bi.

Basque arri, harri "rock", harriaga "rocky place", may be related to Niger-Congo ka, kan, kal "coal, charcoal", Ewe aka "coal", Nupe eka "charcoal", Gbe naka "charcoal", Anang nkan "charcoal", Djarawa kal "charcoal", Kamuku kala "coal", Bulong ninkar "coal", with k > h > #. Bantu -kala "charcoal". Likewise haitz "rock", "stone", haitzaki "stone" etc. For the suffix -ri see Westermann (1927 : 108, 183 and 184, 3).

The words aitzur "hoe", aitzkora "axe", aizto (Roncalese) "knife" are allegedly from the same source, with loss of h-. This interpretation, which goes back to Lucien Bonaparte, is rightly questioned by Trask (1997 : 289), particularly for the words haiztur "tongs, forceps", and haiztur "shears", which could hardly be made of stone. Trask (1997 : 142-143) notes that in Basque intervocalic lenis l was converted to tap r, which explains why Basque has a medial r in this word.

Basque gizon "man" (Aquitanian CISON) may be connected with Ngizim (W. Chadic of Nigeria) gimsik, gimsak "man, men". A loanword? The Basque g should come from a Niger-Congo *gw. Mande has an initial k in most words for "man", Landoro (Koelle) hinga & hi, Mende hindo but Mano, Gio, Dewoi, Bassa, show a g (Mano go "man", Bassa ga "man"). Logone (Chadic) has g3n3m, g3nam "woman, women". Women seem to have originally had a special prominence among the Basques.

The Chadic languages lie spatially between Berber and Niger-Congo, which we consider to be related to Basque. It is possible that Basque comes from this general focal area.  Niger-Congo is now thought to have early connections with Central Sudanic. The ancestor of Basque may have originated near Lake Chad, on the black side of the border, so to speak.

Basque biotz "heart", Niger-Congo bi "female breast", Nupe ebe "breast", Eafeng ebi "breast", Akurura ebi "breast", Mossi bihli "female breast", bisum "milk", Barba bei "female breast", Kissi bir "female breast", Kussassi bisim "milk". For the suffix see Westermann (1927 : 109, 184, 6 and 7 -t, -s).

Basque belar "grass", Niger-Congo pi "grass", Guang ipi "grass for roofing", Alagiang efi "grass", Bassa pi "grass", Dewoi pi "grass", Grebo pidi "grass",

Bowili lipili "roofing grass", Likpe kofimi "roofing grass". The la is an article, with added -r (Westermann 1927 : 108, 184, 3). The original i vowel has changed to e.

Basque pareta "wall", Niger-Congo Togo R. language, Adele opore "wall", Guang opol "wall". The o vowel is probably due to influence from the labial, i.e. it was originally an a. Added suffixes -re, -ta, Westermann (1927 : 108 - 109).

Basque seme "son", Niger-Congo gia, ghia "blood", Ewe kadze "blood", (Koelle ekagie), Tschi boga "blood", Guang obudza "blood", Yoruba edzhe "blood", Nupe edzha, egia "blood", Gbari aga, agya "blood". The -me is from -mbi, Niger-Congo bi "child".

Basque senar "husband", Likpe osani "man", Santrokofi osa "person", Adele osan "husband", Akpafu ora (t > s > r) "husband". Compare Niger-Congo ta "father", Ewe ata "father" etc. Basque has a > e here. Added -na "mother" + -r.

Basque ar "male", Niger-Congo kua "man, slave", often with a labialised stop, Ahlo u-kpa "slave", Tschala o-kpa "man", Guang e-kpa-bi "slave" (with bi "child" added).

Basque mihi, min (< *mini ) "tongue", Niger-Congo lima "tongue", Edo olemi "tongue", Efik edeme "tongue", Mfut derim "tongue", Bidjogo ninume, nunume "tongue", Bantu limi "tongue". [l < m]. Trask posits original initial *b for this word (Trask 1997 : 287). Westermann (1927 : 252) mentions a "secundäre" b in groups IV (Gur) and V (W. Atlantic), but cites Bantu evidence, -limi "tongue", for original initial l.  Assimilation. This word may give an indication of the group(s) of Niger-Congo that Basque belongs to. The change n > h is standard in Basque. Compare Avatime, Nyangbo mini "to taste, sample".

Basque hatz "finger", Niger-Congo ka "finger, hand, arm", Yoruba ika "finger", aka "arm", Igara ika "arm", Igbo aka "hand", Ahlo ika "hand", Biafida -ka "five", Padjade nka "five". Suffixes -t, -s.

Basque su "fire", Niger-Congo (Eastern Kru) kosu "fire". [Basque also has egur "firewood" (Trask 1997 : 311)]. Compare Adampe ezo "fire", Anfue itso "fire", Hwida ozo "fire", Dahome zo "fire", Mahi uzo  "fire", all from Koelle, and from the same general area.

Basque lur "earth", Niger-Congo lu "powder, ash, dust", Ewe du "powder", Guang odu, olu "powder", Yoruba luru "powder", Avatime alu "powder", Akpafu kudu "powder", Kpossi udu "powder", Gola edun "ash", Kpelle du "ash, dust". Suffix -r.

Basque mu "cow", Niger-Congo lu "cow", Ewe lu "cow, antelope", Ani lu "cow", Abe lu "cow", Lefana kedu "crested antelope", Avatime olu "antelope, big, black", Ahlo ulu "cow, antelope", Santrokofi dufu "crested antelope", Gio du "cow". The m of Basque is due to a lost nasal prefix. 

Basque Jainko "god", Niger-Congo gia "to be" (in a place)?, Lefana ya, a-ya "god", Santrokofi ya "god", Akpafu ya "god", Guang, Brong, Yoruba, Animere, Kposso, etc. ya "to be". Basque also has a form Jinko, which recalls Niger-Congo gi "to be" (in a place), Nupe yi, Temne yi "to be". The -ko is from gbo "heaven, god", Logba na-gbo "heaven", Ga no-nmo "heaven, god", from an original Niger-Congo labiovelar *gw



The Africanist Mukarovsky includes morphology in his comparisons of Basque with Berber, comparing the Basque male suffix -k (< ga) and female suffix -n (< na) with Berber suffixes of similar meaning, -ak and -am. Berber is thought to have been spoken on the Canary Islands (Guanche). We think the Basques went to Spain via the Canary Islands.

Morphology is the acid proof of language relationship. Thus Hayward (in Heine and Nurse 2000 : 87) declares: "Morphology is the surest test of genetic relationship". Languages rarely if ever borrow morphology. So Mukarovsky is on the right track in using morphology here, and in looking for systematic correspondences.



A better comparison here might be with Niger-Congo na "mother" etc. for the feminine. Westermann cites forty-four reflexes for this root in Niger-Congo.



For the masculine, Niger-Congo has the word ga "big, large", but also "leader, hero, headman, king" (fourteen reflexes in Westermann). But ga should give Basque h/#. Perhaps from *kwa  "man".

We can also explain our Basque "feminine form" in our sentence "I have seen the house" with inserted na (dut but fem. dinat). Compare Basque andre "of woman", which may incorporate the same element na, with loss of initial n-. It is said to be from Celtic however.

Niger-Congo lacked grammatical gender, but added words for "man" and "woman" if gender was required. So our explanation fits.



The Basque plural in ‑k is not original and can only be used after the definite article, hence the form  -ak. We derive this from Niger-Congo gua/gwa "much". This derivation is supported by the meaning "all", which is often given for this plural. The simple word without ‑ak is also used for singular and plural indefinitely: gizon "man" but also "men". The original plural form was -ri, as in Etruscan (> r), as can be seen from the numerals. This r often precedes this -ak  plural. The theory that this form comes from earlier *g appears right (Trask 1997 : 200), to judge from our derivation: gw > g > k.



In general the numerous Basque "cases" appear to be made by adding postpositive Niger-Congo grammatical roots which were originally full forms, often verbs. Trask (1997 : 201) regards only a few of the Basque cases as original: Absolutive, Ergative, Instrumental, Dative. See also Lafon's exposition (1960 : 77) of cases, which differs from that of Trask. Where a case ends in a consonant Lafon (1960 : 89) cites word final accents in Souletin, showing that a vowel has been lost.

Many cases are based on what have become opaque agglutinative formations. Often a plausible Niger-Congo identification can be made, but not always with certainty. Authorities differ as to the number of cases and their interpretation. The names of the cases are differently given by different grammarians. We have used the terminology followed by Trask (1997). Lafon (1960 : 77), for example, has quite a different terminology.



This is a formal zero. It is used for the subject of an intransitive verb, the direct object of an intransitive verb, the complement of a copular verb, and as a vocative. Such usages are characteristic of ergative systems.



Nouns have an ergative suffix -k used to indicate the agent "with a transitive verb" (Trask 1997 : 92); but thought by many (not Trask) to be an agent phrase used with a "passive" verb. The Absolutive case (zero marker) is used as the subject of intransitive verbs, but as the direct object of a transitive verb. This is known as an ergative system. The -k  already occurs in the 9th century Emilian Glosses, and is clearly old. It probably comes from Niger-Congo gu/ghw "he" (Westermann 1927 : 224, with thirty-two reflexes). In Western Sudanic it is normally reduced to wo, we, u, o. The original vowel was weak.



The genitive (possessive) follows the noun in most Niger-Congo languages (Hawkins 1983 : 290). Basque has a genitive in added -re, -ne, -ren, -en, -len, -lon. These Basque genitives are probably from Niger-Congo na "to be, to remain" or Niger-Congo ni "in, inside, by, to", with an added article ra, re, la, le. Etruscan also uses the Niger-Congo article as a pseudo-genitive.


Dative: Basque ri, i, Niger-Congo ki "give"; Avatime (Togo R.) ki is the basis (Westermann 1927 : 112, 19).


Benefactive: Basque Genitive ending + tzat. Niger-Congo ta, tsa "possession", ta "give". Compare Lafon (1960 : 89).


Comitative: Basque gaz, kaz (pl), kin. "with, in the company of". Niger-Congo ga "side", "by", also gi "two".


Instrumental: Basque z < t. "instrument". Niger-Congo ti, tu "to take", or ti "open, close".


Locative: Basque tan. "in, on, at". Niger-Congo ta "top, on" + ni "inside".


Ablative: Basque tatik. "from" + particle k. (< gi "go"). "out of, away from". Niger-Congo ta "take" + ti "to take" or ? ti "middle, between".


Allative: Basque tara. "towards". Niger-Congo ta "on" + la "under, below, earth".


Directional: Basque tarantz. "for". Niger-Congo ta + ra + ti "middle".


Terminative: Basque taraino. "as far as, up to, until". Niger-Congo ta + ra + i + no. The i has a lost consonant, probably l/r/d. The no is Niger-Congo na, no "lie, remain, sit".


Destinative: Basque -(r)ako. (Some northern dialects have -rakotz.) Used for the inanimate entity towards which some action is directed. The essence of the case is Niger-Congo ra "towards" as in the directional. The ko is probably from Niger-Congo gua, gbo "gate, go".


Trask (1997 : 93-94) mentions also a partitive which is "not properly a case", but "like a determiner" (= Niger-Congo gi "this"), a locative genitive, really the adjectival -ko, an essive/translative/prolative from Niger-Congo gia "to go" + ti "to take", and a causal -engatik "because of", which is really a postposition. His reasons for not regarding these as cases are stated on page 94.


Position of Cases

Basque adds case after the plural: "Det 2 - number - case" (Trask 1997 : 89), just like Etruscan: clen-śi "to the son", clen-ar-śi "to the sons". Etruscan is also Niger-Congo.



This is formed with the suffix -ko (Trask 1997 : 100), which "can be added to any kind of adverbial phrase…to produce a complex adjectival modifier" (slightly abridged).

Basque atzo "yesterday", atzoko "yesterday's", hemen "here", hemengo jendea "the people here", mendian "on the mountain", mendiko etxeak "the houses on the mountain", Bilbon "in Bilbao", Bilboko kaleak "the streets of Bilbao", etc.

The same formant can be added to a N-bar (Trask 1997 : 102) "to produce a modifier". Bihotz on "a good heart", bihotz oneko neska bat "a good-hearted girl", hortz bi "two teeth", hortz biko sardea "a two-pronged pitchfork".

Addition to verbs is rare: balitz "if it were", balizko "hypocritical" (Trask 1997 : 102).

This element -ko can be identified as Niger-Congo gua "much" (Westermann 1927 : 217), and is the same as the adjectival formant in Etruscan. Ga gbe "much", Nupe gbagba "very, exceedingly", Temne gba "much", Adjukru gba "big", Ewe gbogbo "much", Yoruba gbogbo "all, every", Sussu gbo "much". The labiovelar gives Basque k as is usual. The a-vowel often changes to o in Niger-Congo. Niger-Congo also has forms with e: Bulom gbe "much", Temne gben "very much", Sussu gbe "much". Basque has somewhat reinterpreted this suffix.



Basque beltz "black", Niger-Congo bi "black", Tschi biri "to be black", Guang bir, biri "black", Beri biri "black", Gba bila "black", Temne biti "to be black" etc. Compare Niger-Congo pi (+ l ) "black", Mfut fin "black", Tara pini "black", Soninke fin "quite black", Bulom pit "black", Bamana fin "black", Djula fi "black", Bozo pi "black", Vai fi "darkness", Bantu pi "darkness". The contrast between voiced and voiceless stops in Niger-Congo is being lost. This is the best identification among the colour words.

Basque zuri "white", Niger-Congo gu "cotton tree"; Azkue (1923 - 26 : 46) links this to zur "wood". Another possibility is pu "white", Nyangbo, Adele fu "white, to be white", with change of p > f > s.

Basque gorri "red", Niger-Congo gu "kill", Delo wo "dead", connected with gia "blood" [or guia "sun" (compare B & G gori "hot")]. Compare however Niger-Congo gian "red" (like blood), Ewe dza "red", Grebo djao, djoo "red", Igbo dzha "red", Nyangbo dza "to be red".

Basque hori "yellow", Niger-Congo kua "die Farbe verlieren", Tschi kwa "to wear off", Ga kwa "to lose colour", Grebo kpro "to fade".

Basque urdin "blue", but originally "grey", "white", "black", Niger-Congo lu, du "dust, ash, powder", Ewe du "powder", Guang odu, olu, kuduri "powder", Yoruba ruru, luru "powder", Avatime alu "powder", Akpafu kudu "powder", Kposso udu "powder", Gola edun "ash", Kpelle du "ash". Common sound change u > i in Basque.


Other adjectives

Basque haundi, handi "big", Niger-Congo gan "big", Ewe ga "big", Ga gaga "long, high", Ani gagraga "big", Yoruba agaga "the topmost point", Nupe gaga "highest point", Bamana gana "grand", Gagu gaga "big", Edo oga "headman". In Basque, original Niger-Congo g develops like k.

Basque huts "empty" (used for "zero"), Niger-Congo pu, pul, "empty", Ewe ful "empty", Tschi huhuw "empty", Guang fulu "empty", Grebo pulu, pudu "to be empty", Yoruba ofurfu "emptiness", Igbo afulu "empty", Efik ifu "vain, worthless", Avatime pluplu "empty", Temne fu "in vain", Sussu fu "nothing". Compare pol "empty".  The Tschi form with an initial h explains the initial h of Basque.

Basque eder "beautiful", Niger-Congo lia, Togo R. languages Lefana dia "to be beautiful", Santrokofi le "to be beautiful", Akpafu ile "beautiful", Guang ale "beautiful", Tschi de "good, pleasant".

Adjectives are normally but not always postposed in Basque (Trask 1997 : 210) as in Niger-Congo. Modifiers in -ko are preposed however, as are formations with -dun "having" and ethnonymic -(t)ar, adjectives of nationality. Likewise numerals. Adjectives were originally not distinguished from nouns (Trask 1997 : 210). Niger-Congo has practically no specialised adjectives.

Basque adverbs in -ki and -to may be from Niger-Congo gi "to be full" (compare English "tear-ful", "beauti-ful" etc.) and tu, to "to work".



The Basque verb is regarded as a formidable challenge due to its multiplicity of forms which allow prefixation, infixation and (inflectional) suffixation. Like the Niger-Congo verb, it differs from the noun in that the finite verb is not usually prefixed. This distinction is fundamental.

There is a significant exception however. "Virtually all ancient verbs show a prefix *e- in all their non-finite forms. This appears today variously as e-, i-, j- or "zero" (Trask 1997 : 211). Trask considers this prefix "originally derived a verbal noun from a verbal root". This prefix can be seen in the "radical", the stem of the verb functioning as a free-form (as in ikus, radical of ikusi "see").

The copula izan has an initial vowel in the infinitive, likewise egon "wait, remain", used as an alternative copula. The verb iduki "have" also shows this feature, as does ibili "to walk" and ekarri "to bring" and jarri "to sit down". Others: egin "do, make", egon "stay, wait", eman "give", esan "say". Finite forms of these verbs however lack this initial vowel.

Now, in Niger-Congo a prefix i- is attached to certain verbs for purposes of nominalisation: ka "to say", but Efik iko "a word, an utterance", ku "to kill, to die", Kpossi iku "death", Bowili eku "death", "to scoop, bale water", Igbo iku "drawing water", li "to eat", Igbo ili "eating". Compare Bantu ikala "to remain" from kal "remain, sit". The infinitive is a type of verbal noun. This, we think, is the reason why certain Basque infinitives are prefixed.

On Yoruba Westermann says "i- bildet vorwiegend den Infinitiv" (Westermann 1927 : 76): ba "treffen", i-ba "das Treffen",  fo "reinigen", i-fo "das Reinigen", gba "empfangen", i-gba "das Empfangen" etc.

Words with initial j in the infinitive originally had the i- prefix. Thus joan "to go",  jo "to strike" etc., once had a labiovelar consonant preceded by this prefix. This combination issued in  j (differently sounded [y, h, χ] in different dialects). The imperative oa  lacks the j. Elsewhere j is from i + d as in jan "to eat" < Niger-Congo dia "eat". This j is "strictly word initial" (Trask 1997 : 88).


Verbal Derivational Suffixes

The Basque verb, if we assume it is basically monosyllabic, apparently manifests a range of derivative suffixes of unknown meaning: ikusi, ikarri, joan, ibili, hautsi, entzun, jakin, erre, etorri, ebaki, moztu, zulatu, edan, jan, jausi, isuri, eman etc. Westermann (1927 : 108, 12, 184) lists suffixes used with the verb in Togo R. as follows: -ni, -mi, -li, -ri, -bi, -k-, -t-, -si, -s-, -f-.  He considers the identity and meaning of these elements unknown. Basque (Trask 1997 : 88) permits only t, k, ts, tz, tx, s, z, x, n, r, l in final position, with neutralisation of the nasals and no f. This, all things considered, gives a reasonably good match with most of  the Basque verbal "suffixes". We have already argued that Basque is a Togo R. language, and this confirms it. Togo R. verbs tend to be disyllabic (Westermann 1927 : 108, 12, 184).

The Basque Present Tense is marked by da-, -a-, of which we regard d as an old pronoun, and the vowel a as the original tense marker (Trask 1997: 223).  This a vowel is also typical of present tenses in Niger-Congo languages (Stapleton 1993: 151, Table 434).   Examples include Poto kala, Ngombe pala, "love" (simple present). We note that Greenberg has identified Ablaut vowel a as a marker of the present in Proto-Afro-Asiatic.

Take the verb "come". Hualde & Urbina eds., Grammar of Basque (2003 : 234) gives 1s nator "I come", 2sf hator, 3s dator 1p gatoz 2s zatoz 2p zatozte 3p datoz. For "walk" he has nabil, habil, dabil etc. Westermann's word for "come" in Niger-Congo is na (fifteen reflexes), for "go" also na (twenty-seven reflexes). This word has a weak initial nasal. His word for "I" is na (seventeen reflexes), ni (fifteen reflexes). Allowing for changes to the initial consonant conditioned by the personal pronoun, we can set up Basque : Niger-Congo protowords ni-na, ki-na > hi-na, di-na etc. as the root. Heine & Nurse have reconstructed si for the third person in Proto-Afro-Asiatic. An earlier dental is likely here (they have ta as fem. in Chadic), whence Basque d.

Curiously the word for "be" is likewise na in Niger-Congo (seventeen reflexes in Westermann), also with weak initial nasal. It also means "sit, lie, remain". Basque has a verb nago, hago, dago, which means "be" but also "stay". This can hardly be a coincidence. It is too remarkable. It constitutes what Meillet (1954 : 3) calls the significant "détail singulier". Ultimately comparative linguistics relies on data that cannot be accidental, given the arbitrary nature of the linguistic sign (Meillet 1954 : 2).

The forms (Trask 213, Table 83) naiz, haiz, daiz "I am", "thou art", "he etc. is" have equivalent pasts: nintzen, hintzen "I was", "thou wert", "he etc. was". One is reminded here of the Ablaut in Arabic Perfects katab, katabt versus Arabic Imperfects 'aktib, tiktib etc. As Trask says (p. 213), the forms of the verb "to be" are especially irregular in many languages. Here we appear to have ancient a/i apophony derived from Afro-Asiatic. Greenberg (1952 JAOS) identified a "present stem" internal ablaut to a (Heine & Nurse 2000 : 91), which may be relevant here.

Corresponding Past Tenses in Basque show e-, ze, which Trask describes as "fossilised bits of morphology". Lafon (1960 : 90) relates the vowel e/i to the unreal "no real". Congolese dialects (Stapleton 1903: 151) show a similar tendency to change the vowel to indicate past tense: Poto kali, Ngombe pali "loved". There are also formations with added -ki (kala-ki, pala-ki) or added -ka. Some forms have a prefix te- (te-kunda "loved" in Soko).

The ending of the Past Tense in Basque is invariably the suffix -r(e)n. This follows all other morphs in a finite form (Trask 1997 : 224): nator "I'm coming", nentorren "I was coming" and is of unknown origin. We derive this in parallel to Niger-Congo gia "to go" from Niger-Congo na "to come", Ahlo na "to come" (used as an auxilliary), Djula ne "to come", Nyangbo na "to come". Johnston (1911 : 362) treats a as the normal terminal vowel in Bantu, but e as the subjunctive formant. It is possible that this form in e represents an ancient Irrealis, converted for use as a past.

Trask (1997 : 224) notes the existence of a morph -ke, with a variant -te, -teke "which are explicitly marked for potential mood, though not exclusively so": etor daiteke "he can come", etor liteke "he would be able to come", etor sitekeen "he would have been able to come".

The Potential (Hualde & Urbina 2003 : 217) forms include this affix -ke, which I derive from Niger-Congo gi, gia "to be" (in a place), Grebo yi "to be" (auxillary verb), Tschi ye "to be", Ga ye "to be", Gurma ye "to be", Malinke ye "to be in a place". The Niger-Congo labiovelar becomes not a y but a k in Basque: daiteke "he she, it can be".

The origin of this -ke "is completely mysterious", says Trask. But in the past this morph also marked futurity: noa "I'm going", noake "I'm going to go". This archaic future appears to represent an extension of potentiality to futurity (Trask 1997 : 225). In fact Kele has a-lembe-ke "loved", e-lembe-ke "will love" (future indefinite) as can be seen from Stapleton's (1903) Table 434 on page 151. This early system aligned the Future with the Past, as Unreal (not actual).

Compare Basque dakidake (Present) with Basque zekidakeen (past) and lekidake (Potential) in Hualde & Urbina eds. (2003 : 218). Latin and Greek have subjunctives in e and optatives in i (Latin siem, sies, siet) which may be due ultimately to Niger-Congo substrate.

Westermann (1927 : 137) reports comments by Fisch on the conjugation of the verb in Dagomba (Gur) which include a formant in da (vorgestrige Imperfect, fernere Vergangenheit) and in re (Progressiv), also a Future in ni. Similar forms existed in Tem. These procedures are very ancient and much remodelling has taken place in Basque. The evidence points to Niger-Congo as the source.

Basque ancient verbs form their perfective participle by adding a corresponding suffix -i to the radical (Trask 1997 : 212): ikusi "see", ekarri "bring", ibili "go about",  jarri "sit down". An identical i can be used to derive participles from nouns and adjectives: itzal "shade", itzali "to obscure", hauts "dust, powder", hautsi "to break", igun "repugnance", iguni "to detest", zorrotz "sharp", zorrotzi "to sharpen".

This suffix also occurs in Niger-Congo with "keine erkennbare Funktion": wa-i "kratzen", ku-i "schöpfen", fo-i "ein Zeichen geben" (Westermann 1927 : 93). The examples are from Benue-Cross, and include plural forms of the verb: duat "schlagen", duari (plural form). It will be remembered that i (from hi) forms plurals of nouns in Ga and Kunama (Westermann 1911 : 45-46), but appears in Nuba without changing the meaning. Note however that Johnston (1911 : 362) mentions an i which gives "a negative sense in certain tenses" as a root termination in Bantu. Compare the Swahili negative auxilliary si. This negative auxilliary is thought by Torrend (1891 : 232) to be derived from the verb sia "to leave, to avoid". Presumably the same word as PWS ki "to abstain".

Western dialects use added -ta + adjectival -ko for the perfective participle, with subsequent addition of the article, e.g. Ibaiak kutsatutakoak dira "The rivers are polluted." The final -k is the plural, with preceding -a article. The ta here is Niger-Congo ta "to sit, be accustomed", used as "Habitual" (Westermann 1927 : 282), with reflexes in Guang, Ga, Edo, Ahlo, Santrokofi, Animere, etc.

There is also a Basque suffix -te (variant forms -tze, tza etc. exist) which forms nouns of duration or abundance (Trask 1997 : 215): euri "rain", eurite "rainy spell",  jende "people",  jendetze "crowd". I relate this to Kwa -te (Westermann 1927 : 59): bo "närrisch sein", bobo-te "närrische Dinge", which comes from de "thing". Compare de "thing", de-te "anything". It is a type of intensive plural. Compare Westermann (1927 : 110, 11), "Mz von Dingen" in Nyangbo. Also compare the Togo R. (Nyangbo) "perfect" suffix te (Westermann 1927 : 112, 20).

These features appear very ancient and are not fully understood in either Niger-Congo or Basque. They constitute a very significant morphological detail.

Also significant is the question of the Basque "dialogue conjugations" in the verb. Basque verbs may incorporate special pronouns for number, gender, person of the addressee, - along with the subject and object pronouns. (Such familiar speech is called tutoiment in French.) Incorporation of extra pronouns in the verb is a feature of certain Niger-Congo languages, Migili (S. Plateau) and Kente (Jukunoid) for example. It also occurs in Chadic, where it is called the ICP "Intransitive Copy Pronoun" construction. Intransitive verbs fill the object pronoun slot with a pronominal element (Gerhardt in Bendor-Samuel 1989 : 373 - 4). Such "surplus" pronouns may provide the basis for the Basque dialogue construction.

Basque verbs have a "fully productive causative suffix" (Trask 1997 : 113) -erazi, -arazi, (dialect -erazo, -arazo), e.g. etorri da "came home", etorrerazi du "made to come home" (3rd person singular). This has a parallel in Bantu causatives: Kongo -isa, -esa (according to verb class), Bangi, Poto -isa, Lolo, Ngombe -eza (Ngombe z approximates to dz), Kele -esa, -ese  (according to word class), Swahili -sha, -za (Stapleton 1903 : 212 - 213).

There is also "an ancient causative prefix -ra-" (Trask 1997 : 114) seen in a few common verbs: joan "to go", eroan, eraman "to take away", ibili "be active", erabili "put in motion"; some are semantically anomalous, ekarri "bring", erakarri "attract",  jantzi "get dressed", erantzi "get undressed". The origin of this "prefix" is Niger-Congo la, da "to make" (Westermann 1927 : 246), Djula la, Malinke, Sussu, Boko da (all Mande languages), Guang lala. Chwana (Bantu) has causative -ra (Torrend 1891 : 278).

Trask's (1997 : 217) "agent suffixes" in -le, (erosi "buy", erosle "buyer") and  -n (most often there is an -ile variant) are related to Bantu formations (Torrend 1891 : 278) in -la or -na which are old Causatives. Trask notes that Bizkaian dialect has forms in la. Trask regards this form  as original. This suffix is really the Niger-Congo article/pronoun (Westermann 1927 : 247), with meanings in Togo R. of "he, this, Definite article, 'bildet nomen agentis' [Adele], possessor".

The Basque suffix -tzaile, (tza) denoting "agent" is from Niger-Congo (Bantu) -zia, za, Chwana tsa given by Torrend (1891 : 278) as causatives. The suffixes -tze, -te which form Basque gerunds (Trask 1997 : 215) come from the same source. (Compare the Latin gerund meaning "requiring to be….)

The Basque verb ibili "go around" is really a Niger-Congo applicative (Torrend 1891 : 276) meaning "for, to, into, around". Torrend gives -ila, -ela, -ira, -era, -ia as forms of this suffix. The original Niger-Congo root is ba, be, bia (Westermann 1927 : 203 with 209) meaning "be in a place", but also "come". We assume ba > be > bi (or primary bi).

A summary of Bantu causatives, reciprocals, and verb forms incorporating "some other person" (us. = the Niger-Congo article-pronoun), all expressed by suffixes, as in Basque, can be found in Johnston's article "Bantu Languages" in EB (1911 : 362).

The conditional (Hualde & Urbina eds. 2003 : 216) in Basque has the prefix ba- meaning "if". This is the Niger-Congo root ba "to be".  E.g. balira "if they were", etorriko balira "if they came". Compare the English conditional expression "be he alive or be he dead…".

The verbs of motion are interesting. Basque uses different verbs for "going in, out, up or down" (Trask 1997 : 295). Niger-Congo (Stapleton 1903 : Congolese vocabulary) makes a similar distinction between "come" kwiza, ya, "arrive" kita, kuma, luaka, "come back" luta, simba, bula, rudi, "come from" tuka, lima, uta. For "go" Stapleton lists ke, kwenda, kenda etc. In other words this semantic field is segmented in a similar way in both languages.

The Basque words are: sartu "go in, enter",  jalgi, elki, urten, irten, ater "go out", igan "climb, ascend",  jaitsi, beheratu "descend", kendu "to remove oneself", the last of which is obviously connected with Lolo kenda, Ngala, Kele kende "go". Basque jalgi relates to Bangi, Lolo, Ngala, Poto, Kele ya, Swahili ja. For Basque jalgi, elki compare Lolo kita, Swahili fita "arrive".

Basque sartu "go in", Niger-Congo ta "to go", Likpe to "to go", Avatime ta "to go", Numu ta "to go", Vai ta "to go", Gbe ta "to go", Boko ta "to go", Bantu tamba "to go" (with ba "go"). Basque amaitu "to go", may be from Niger-Congo ma "to finish, to leave off", or from Niger-Congo ba "come".

Basque egokitu "to go" and egoki izan "to go" are related to Niger-Congo kua "to go" and kua "leg, foot", the latter of which sometimes has a voiced labiovelar which would give Basque k. The ki may be from Niger-Congo gi "to go", Bantu γia "to go", where Meinhof's γ represents a labiovelar.

Basque hil "kill" or "die", Niger-Congo ku "kill" or "die", Ewe ku "to die", ku "death",

Tschi ku-m, ku "to kill", Afema ku "to kill", Mekabo ku "to kill", Gwa ku "to die", Kyama ku "to die", Newole ku "to die", Grebo ku "demon, departed spirit", ku tu "dead body", kw-e "to die", Yoruba ku "to die", Okpoto ku "to die", Igara ku "to die", Nupe eku "carcass", Igbo ku "to die", Edo ku "to kill", Dujku ki "death", Bowili eku "death", Kpossi iku "death", Lefana kpi "to kill", Santrokofi kpi "to kill", Akpafu kpi "to kill", Mossi kia "to die", Dagomba kpi "to kill", Dagarti, Birifo, Gba, Lobi, Djan k'i (< kui) "to die", Tobote kpi "to kill", Gurma ukpi "death".  Basque also has an adjective hil "dead". The use of the same word for "kill" and "death" in both Basque and Niger-Congo is significant, as is the occurrence of the vowel i in Westermann's II Benue-Cross (only Djuku), III Togo Remnant, and IV Gur. The phonetic development with i is most well-established in Gur dialects, which may be an indication of where Basque split off from Niger-Congo. Compare however the following example.

Basque lo (egin etc.) "sleep", Niger-Congo la "to lie, to sleep", Tschi da "to lie, to sleep", Guang da "to lie", Igbo la "to sleep", Mfut da "sleep", Yala la "sleep", Likpe da "to sleep", Lefana le "to sleep", Kpossi le "to sleep", Bulom lol  "to sleep", Krim lo "to sleep", Kissi lo "to sleep". Bulom, Krim, and Kissi belong to the West Atlantic Group and are therefore particularly relevant. This word appears not to occur in Westermann's Group VI, Mande (Mandingo). 

Basque esan "to say", Niger-Congo ta "to say, to tell", Ewe ta "to tell", Nupe ta "to tell", Igbo ta "to talk, to tell story", ita "proverb story", Edo ta "to say", Effik tan "to speak", Dagomba tohe, toγe "to speak", Atjulo ta "to say", tare "to speak", otare "word", Tschala tage "to speak", Temne tan "to promise, tell", Bantu teta "to speak". The change of t to s is common in Niger-Congo.

Basque esan also means "to give", Niger-Congo ta "to give", Alaging ta "to give", Yoruba ta lo re "to give a gift", Logba ta "to give", Likpe te "to give", Lefana ta, to "to give", Santrokofi ta "to give". As in the previous example gives s. This word also means "to leave" in Temne, Ewe etc. It may be related to the Niger-Congo word for "hand" ta, sa and to the Niger-Congo word "to be accustomed".

Basque eman also means "to give", however. This may be related to Niger-Congo na "to give", Ewe na "to give", Guang ne "to give", Obutu na "to give", Kukuruku na "to give", Tivi na "to give" etc. There is a problem regarding Niger-Congo m, which tends to be secondary. Pre-Basque is thought to have had N  (fortis) and n (lenis). That is n > m.

Basque eduki "to have", Niger-Congo tu, ti "to take", Guang tu, tsu "to take", Avatime tu, tuku "to take", Nyangbo tuku "to take", Santrokofi tuka "to take", Ahlo tu "to take", Mossi tukia "to take", Adjukru tsu "to take", Bantu tuala "to take".

Note that PWN THU "carry" (on head) may be related. Likpe has tsi "to take" with an s as in Guang and Adjukru. This suggests an original Niger-Congo aspirate, which appears to be responsible for the d of Basque.

Basque aditu "to hear", Niger-Congo tie "to hear", Ewe se "to hear", Tschi tie "to hear", also tse "to hear", Zema de "to hear", Mekabo te "to hear", Abe tye "to hear", Abikan se "to hear". Compare also PWN THUI, THU  "ear" . The Niger-Congo sound *th appears as Basque d, as in the previous example.

Basque words for "eat and drink" are jan and edari respectively. These both come from Niger-Congo lia, dia "to eat", which is also used of drinking. In Basque the two words have diverged and developed separately.

For jan "eat", compare Kwa dji "to eat", Yoruba and Okpoto dzhe "to eat", Kpossi dji "to eat", Tobote dji "to eat", Gurma dje "to eat", Gola dze "to eat", Adjukru dzhi "to eat", Bozo dye "to eat". These words show the basis for the development di > dj > j as in Basque.

The word for "drink", edari in Basque, shows a different development with suppression of the i vowel after d. Compare Efik dia "eat", udia "food", Anang dia "to eat", Akpafu de "to eat" Kussasi de "to eat", Tschala adia "victuals", Gurma dje "to eat", Temne dia "to eat with", Bola de "to eat", Bozo dye "to eat". There may be influence from Niger-Congo lum "bite", Guang don "to bite", Animere do "to bite", Susssu do "to eat". The ending ri occurs in a number of Basque verbs, e.g. erori "to fall".

Basque ekarri "to bring" is from Niger-Congo gua "hand", used of actions of the hand, Yoruba, Igbo, Okpoto wo "arm, hand", Nupe egwa "arm", Bowili awoe "hand", Kposso uwo "arm".

Basque izan "to be, to exist", Niger-Congo li, di, ri "to be", Ewe di "to be", Ga dzhi "to be", Guang dzi "to be", Yoruba ri "to be, to have", Igbo ri, di "to be". For the a vowel, compare Niger-Congo le "to be" (in a place), Ewe le "to be" (in a place) etc., with e from earlier ia.  But influence from na "to be" is also possible. Note however that Niger-Congo gi "to be" (in a place), Grebo yi "to be", Nupe yi "to be", Temne yi "to be", and Niger-Congo gia "to be" (in a place), Yoruba ya "to be", Animere ya "to be in a place", must be related and give a better phonetic: Niger-Congo gi > Basque z.

For Niger-Congo g (a labiovelar) giving Basque z, compare Basque zapi "clothes" with Niger-Congo ga "to put on clothes", Guang wa "to put on clothes", Yoruba wo "to put on clothes", Igbo wa "to put on (cloth)", Temne won "to put on". The pi may be from Niger-Congo pi "grass" (a reference to grass skirts?).

Basque ikuzi "to wash", Niger-Congo gua "to wash", Tschi gware "to bathe, to wash oneself", Kru wa "to wash", Yoruba gwe "to wash" (dialect), Igara gwara "to wash", Barba wale "to wash", Gurma ware "to wash", Gola gwa "to wash oneself", Kpelle wa "to wash oneself", Mende wa "to wash", Sussu γa "to wash".

Basque ikusi "to see", and Basque so "look", Niger-Congo kua "to see" (Westermann 1911 : 154), Ewe kpo "to see", Tschi (hu, hwe, fwwe "to see"), Ga kwe "to see", Efik (kut "to see") [a > e], with ku (labiovelar) > k [> s and ua > o] in so ?.  PWN has KHUP "eyelid". An original labiovelar here.

Basque eman "to bear", umea egin "to bear", Niger-Congo mu, muan "to beget, to bear", Ewe mo "to bear", Grebo mwane "to bring forth", Nupe mo "to bear", Gbe ma "to bear", Igbo mua "to bear a child", Edo mua "to bear a child", Efik man "to bear", Ahlo ma "to bear", Akassele ma "to bear", Gurma male "to bear".

Basque erosi "to buy", Niger-Congo la, ra, da "to buy", Yoruba ra "to buy", Igara la "to buy", Avatime da "to buy", Animere la "to buy", nda "market", Mossi da, ra "to buy", Dagomba da, ra "to buy", Kusassi da "to buy", Gurma da "to buy". Niger-Congo a has changed to o.

Basque zenbatu, balio ukan "to count", Niger-Congo ba, bal "to count", Nupe ba "to count, to read", Gbari ba "to count, number", Igbo bwa "to count", Efik bat "to count", Bantu vala "to count".

Basque jo "hit", Niger-Congo gua "to break, rend, smite", usually gba, with labialised stop, Ahlo gbo "to strike". The word has the merged prefix i- > j-.

It is noteworthy that about 30+ items on Swadesh's 100-word list for Basque can be definitely identified as Niger-Congo: all, and, at, big, black, burn, child, come, count, day, die, drink, eat, fat, father, fire, five, flow, foot, four, give, good, grass, hand, head, hit, take, husband, I, if, in, kill, know, leg, man, meat, mother, mouth. Some of these are redundant, kill and die, eat and drink for example. This should mean that Basque is about four thousand years from Togo R.  Note: Swadesh compares words; we use Niger-Congo roots.


Word Order

Basque is now classified as a non-rigid SOV language, but appears to have previously favoured the order SVO as Trask (1997 : 231) points out. This earlier word order would bring it into line with the Togo Remnant and Gur (Westermann & Bryan 1952 : 100-101). The original object order (OV or VO) of Niger-Congo is debateable (Williamson in Bendor-Samuel 1989 : 28).

I consider that the current Basque word order in the verb phrase has been remodelled to match the order gen + noun (rectum + regens) which is also found in Togo Remnant. In general Basque modifiers precede their heads (Trask 1997 : 122), with the exception of the lowest numerals and most determiners. But compare Trask (1997 : 90) for postposed bi "two" in Bizkaian dialect. In the Togo Remnant likewise adjectives and numerals, also demonstrative pronouns follow the noun (Westermann & Bryan 1952 : 100). But the possessive pronoun precedes its noun in Togo R., as in Basque (Trask 1997 : 91) gu-re mendiak "our mountains".

Basque prefers postpositions to prepositions.  So also Ewe, Fante, Ga (Hawkins 1983 : 325), all Kwa languages, and therefore close to Togo R. The nearby Gur languages also prefer postpositions (Hawkins 1983 : 327). So too the Mande group. Bantu and Benue-Congo use prepositions however, as does Berber. See Hawkins for the details.

The evidence of word order therefore supports our argument.



AULESTIA, G. & WHITE, L. (1992), Basque-English, English-Basque Dictionary, University of Nevada Press, Rheno, Nevada.

BERTRANPETIT, J. & CAVALLI-SFORZA, L. (1991), "A genetic reconstruction of the history of the population of the Iberian Peninsula", in Annals of Human Genetics, ps. 51-67.

BONFANTE, G. & L. (2002), The Etruscan Language, Manchester UP, Manchester.

CARR, Raymond ed. (2000), Spain, A History, OUP, Oxford.

CASAS, P. (1992), Discovering Spain, Pavilion Books, London.

CAVALLI-SFORZA, L. (2000), Genes, Peoples and Languages, Allen Lane, The Penguin Press, London.

CAVALLI-SFORZA, L. & F. (1995), The Great Human Diasporas, Addison-Wesley, Reading, NY, Amsterdam etc.

COHEN, M. (1953), "Libyco-berbère" in Meillet, A & Cohen, M. Les Langues du Monde, Vol. I. ps. 156-64.

COLLINS, Roger, (1990), The Basques, Blackwell, Oxford.

CRISTOFANI, M (1978), Etruschi : Cultura e Società, Istituto Geografico De Agostini, Novara.

DALBY, A. (1998), Dictionary of Languages, Bloomsbury, London.

DIAKONOFF, I. (1982), "Hamitic Languages" in EB, 15, vol. 22, ps. 740-748.

ENTWHISTLE, W.J. (1962), The Spanish Language, Faber & Faber, London.

GARLAKE, P. (2002), Early Art and Architecture of Africa, OUP, Oxford.

GEORGIEV, V. (1966), Introduzione alla Storia delle Lingue Indeuropee, Ateneo, Roma.

GAMKRELIDZE, T. (1989), "Caucasian Languages" in EB, 15, vol. 22, ps. 736-738.

GREENBERG, Joseph (1994), "African Languages" in Collier's Encyclopedia, Collier, New York, Toronto, Sydney.

GUIDO, M. (1963), Sardinia, T & H, London.

HAWKINS , J. (1983), Word Order Universals, Academic Press, New York & London.

HEINE, B. & NURSE, D. (2000), African Languages : An Introduction, CUP, Cambridge.

HOCKETT, C. (1958), A Course in Modern Linguistics, MacMillan, New York.

HUALDE, J. & URBINA, J. (2003), A Grammar of Basque, Mouton, De Gruyter, Berlin, N.Y.

HUBSCHMID, J. (1960-), "Toponimia Preromana" in Enciclopedia Linguistica Hispanica, I. ps. 448-493.

JIMÉNEZ, E. (2001), Struggle and Survival of the Pre-Roman Languages of the Iberian Peninsula, Edwin Mellen Press, Lewiston, New York.

JOHNSTON, Sir Harry (1911), "Bantu Languages" in EB, 1912 edition.

JOHNSTON, Sir Harry (1919-22), A Comparative Study of Bantu and Semi-Bantu Languages, Clarendon Press, Oxford.

KAMEN, H. (1973), A Concise History of Spain, Thames & Hudson, London.

KURLANSKY, Mark (1999), The Basque History of the World, Johnathan Cape, London.

LACOMBE, (1952) "Basque" in Meillet, A. & Cohen, M. Les Langues du Monde.

LAFON, R. (1951), "Concordances morphologiques entre le basque et les langues caucasiques" in Word VII, ps. 227-224.

LAFON, R. (1952), "Études basques et caucasiques" in Word VIII, ps. 80-94.

LAFON, R. (1960), "La Lengua Vasca" in Enciclopedia Linguistica Hispanica, ed. M. Alvar et alii, Madrid, 3 vols.

LAPESA, R. (1965), Historia de la Lengua Española , Escalicer, Madrid.

MARUI, A. (1986), The Africans, BBC, London.

MARTINET, A. (1964), Économie des Changements Phonétiques, A. Franke, Berne.

MEILLET, A. (1954), La Méthode Comparative en linguistique Historique, Champion, Paris.

MEILLET, A. (1958), Linguistique Historique et Linguistique Générale, Honore Champion, Paris.

MICHELENA (1969a), "Sobre algunos nombres vascos de parentesco" in Fontes Linguae Vasconum, I, ps. 113-132.

MICHELENA, L. (1989), "Basque language" in EB, 15, vol. 22, ps. 801-803.

MUKAROVSKY, H. (1977), A Study of Western Nigritic, Institut für Ägyptologie und Afrikanistik, Wien.

PEI, M. (1949), The World's Chief Languages, Allen & Unwin, London.

PIDAL, R. Menéndez (1964), Orígenes del español, Espasa-Calpe, Madrid.

PITTAU, M. (1981), La Lingua dei Sardi Nuragici e degli Etruschi, Libreria Dessi, Sassari.

POP, S. (1950), La Dialectologie, II La Dialectologie non Romane, Ducolot, Louvain.

RELETHFORD, J. (2003), Reflections on Our Past, Westview Press, Oxford.

REPSING, K. (1983), "Ainu Language" in Kondonsha Encyclopedia of Japan, Kondonsha, Japan.

ROHLFS, G. (1952), "Sir une couche préromane dans la toponymie de Gascogne et de l'Espagne du nord" in Revista de Filologia Española" 36, ps. 209-256.

RUHLEN, M. (1991), A Guide To The World's Languages, Edward Arnold, London etc.

SAVORY, H. N. (1968), Spain and Portugal, T & H, London.

SCHULTEN, A. (1930), "Die Etrusker in Spanien" in Klio, XXIII.

STAPLETON, W. H. (1903), Comparative Handbook of Congo Languages, Yakusu, Stanley Falls, Congo Independent State [Baptist Missionary Society].

TAGLIAVINI, C. (1962), Le Origini delle Lingue Neolatine, Pàtron, Bologna.

THOMPSON, J. A. (1980), "Tarshish", in IBD, 3, ps. 1517-1519.

TORREND, J. (1891), A Comparative Grammar of The South-African Bantu Languages, Kegan Paul, Trench, Trübner & Co. London.

TOVAR, Antonio, (1957), The Basque Language, U. of Pennsylvania Press, Philadelphia.

TRASK, R. L. (1995), "Origin and Relatives of the Basque Language : Review of the Evidence" in Towards a History of the Basque Language, ed. J. Hualde, J. Lakarra, R. L. Trask, John Benjamins, Amsterdam / Philadelphia.

TRASK, R. L. (1997), The History of Basque, Routledge, London & New York.

TROUBETZKOY, N. (1964), Principes de Phonologie, Klincksieck, Paris.

UHLENBECK, C. (1947), "La langue basque et la linguistique générale" in Lingua I, ps. 59-76.

VÀRVARO, A (1968), Storia, Problemi e Metodi della Linguistica Romanza, Liguori, Napoli.

VIDOS, B. (1959), Manuale di Linguistica Romanza, Olschki, Firenze.

VINCENT, Mary & STRADLING, R. A. (1994), Cultural Atlas of Spain and Portugal, Andromeda, Oxford.

WARTBURG, Von, W. (1967), Évolution et structure de la langue française, A. Francke, S.A. Berne.

WELLARD, J. (1973), The Search for the Etruscans, Nelson & Sons, London.

WESTERMANN, D. (1911), Die Sudansprachen, L. Friederichsen & Co., Hamburg.

WESTERMANN, D. (1927), Die Westlichen Sudansprachen und ihre Beziehungen zum Bantu, de Gruyter, Berlin.

WESTERMANN, D. & BRYAN, M. (1952), The Languages of West Africa, OUP, London.


EB      Encyclopedia Britannica

IBD    Bible Dictionary