Upper Senegal and Niger, Vol. 2

II. — The name of Ghana.

The name of Ghana has been transmitted to us by all the Arab authors without exception in the form Ghânat, becoming Ghânatou in the nominative, Ghânata in the accusative, and Ghânati in the indirect case (by ghayn, alif, noun, and ta-marbuta). The Black people who are aware of it at the present time, from having read it in works written in Arabic, pronounce it Ganata, just as they do for most Arabic words with the same ending (Fatimata, Aïssata, etc.). I use the form Ghana here because it is the most generally employed in Europe; I could have suppressed the letter h, which the Black people do not emphasize and which undoubtedly must not have existed in the indigenous pronunciation of the word, as I suppressed it in the word "Bagana" (1), but I have maintained it for the sole purpose of avoiding a possible confusion with the name of the current village of Gana near Banamba.

Yaqut (2) tells us that Ghana—Ghânatou in the title of the article—is a foreign word whose equivalent he does not know in the Arabic language. Bekri, on the other hand, informs us that ghana was the title given to the kings of the Aoukar, a title which, by extension, had become the name of the city and that of the empire: one said, no doubt, "the city of the ghana, the country of the ghana" or rather, as the article probably did not exist in the language of the indigenous people (3), "the city or the country of Ghana." Be that as it may, the Arab geographers and historians, including Bekri himself, have all given Ghana as the name of a city and that of the State of which this city was the capital.

This word ghana, having no doubt the primitive meaning of "chief" or "king" according to Bekri, certainly did not belong to

(1) The words Ghâna, Gâna or Ghana, with a long a after the g, and Bâghena, Bâghana or Bâgâna, with a long a after the b and a short a after the g, very likely do not have the same origin; their partial resemblance is due no doubt only to a fortuitous coincidence.
(2) Vol. III, page 770, of the Wüstenfeld edition.
(3) If the usual language of Ghana was Soninke, as Barth supposes with much probability, the thing becomes certain.