Dans la littérature ethnographique, le sens des mots religion, magie et sorcellerie reste trés vague. Aucune définition n'a, jusqu'ici, résisté à l'épreuve comparatiste. Longtemps enlisée dans la spéculation sur les origines, l'évolution et la fonction des faits religieux, l'anthropologie se montre actuellement trés prudente à l'égard de toute tentative de théorie générate. L'emploi de méthodes fécondes et l'approfondissement du travail ethnographique ont permis de mieux connaitre les mécanismes mis en oeuvre par le pensee symbolique, mais les concepts de religion, de sacré, de magie, de sorcellerie restent à construire. Le chercheur se trouve done, sur le terrain, privé de référence decisive, qui puisse l'aider à opérer un découpage de la réalité sociale sans négliger une partie de celle-ci.

In 1909 the Rev. J. B. Purvis wrote of the Gisu people of Eastern Uganda that they ‘distinguish themselves as a race apart from others by the name Basani, i.e., men, whilst all men of uncircumcised nations are called Basinde, i.e., boys’ (Purvis, 1909: 271). In modern Uganda the idea of the Gisu as a nation of circumcised men remains as strong as ever. The biennial circumcision ceremonies act both as a focus for such sentiment and as a dramatic display of its power. From the blowing of horns which ushers in the circumcision year to the final aggregation ceremonies during the following December the entire ritual cycle takes about a year. The actual operations are performed in August, in strict order of clan precedence, with the clan where the practice is believed to have originated cutting their boys first. The day after, the circumcisors move to the next clan territory and so on throughout Bugisu. At their height it would be no exaggeration to claim that the festivities involve the entire population of the District– some 500,000 people– from the young children carried along with the circumcision dancing parties to the elderly who are visited as relatives of the novices, approached for guidance on ritual matters or who act as spectators. At the centre are the novices, with the ordeal of circumcision acting not only to validate their own claim to status as adult men but also to demonstrate the values of the entire community.

In this article we shall examine songs of the Rwenzururu movement of the Bakonzo and Baamba people in Western Uganda. We will be concerned with their different uses, as repositories of history and vehicles of exhortation, and also with their origins in time and space. The main interest in this paper, however, is to find explanations for the simultaneous expression of often strongly contrasted moods, viewpoints and styles in Rwenzururu songs.

One of the outstanding features of Akan social organisation emphasised by ethnographers since Rattray's monograph on the Asante (1923: Chap. 2 esp.) is the interlacement of fundamental matriliny with prominent patrilineal traits. With particular reference to the Fante, on whom this paper is based, Christensen (1954: 5, 107) claims, on some wrong premiss though, that his study of 'the role and importance of the paternal line' leads him to conclude that ‘the Fanti manifest a system of double descent’. My own fieldwork (1966-70) on the Fante asafo – the most distinct social manifestation of their patrilineal tendency– indeed confirms the fact that the Fante still lay strong stress on agnatic descent groups, which in Asante and other Akan sub-groups have become more or less ‘obsolete’ (Fortes 1950: 265f, Busia 1954: 207, with reference to ntoro). This notwithstanding, the combination of patrilineal and matrilineal principles of reckoning descent remains a fundamental attribute of Akan culture as a whole. That is to say, the so-called Fante peculiarity is in fact only a slight variation of a general Akan type.

... explanation is no more than exact description which bears its own interpretation (Evans-Pritchard 1973b: 764).

Social anthropologists increasingly question whether or not the translation of culture is possible, at least in the most profound sense (see e.g. Beidelman 1980: 39). If we are as yet uncertain of the essential characteristics of human perception and cognition (cf. Needham 1980: 19) the task of comparison is problematic. Are we to note how ethnographic data reveal indigenous instances of a polythetic mode of classification, or construct rigid dichotomies to sort them out? A related question is whether or not an ethnographic report is valuable because it serves to verify specific theories or because it invites alternative interpretations.