Entre orientalisme et ethnographie: l'itineraire d'un africaniste, 1870-1926.

When Mestizo Logics originally appeared in 1990 (as Logiques metisses) it was welcomed by several historically-minded reviewers working in the anglophone world (Conrad, 1992; Launay, 1989; Robinson, 1992), among them the editor of Stanford's Mestizo Spaces series, in which this English translation now appears (Mudimbe, 1991). French language reaction was mixed (Dans, 1990, welcoming, Digard, 1993, scathing but on dubious grounds) and later anthropologically-minded discussion negative (Austen, 1992; Schilder and Binsbergen, 1993; Burnham, 1996). Since anthropology (or more particularly `ethnological reason') is the foil to Amselle's `mestizo logics', and history (at least rhetorically) his alternative to anthropology, it might be thought that reviewers and commentators were simply aligning with their disciplinary interests. However, anthropological misgivings were founded on more than disciplinary defensiveness: Amsene's critique of anthropology succeeds in being both overstated and conservative. Mestizo Logics is itself something of a metissage: it veers (in descending order of plausibility) between interesting statements about ethnicity in the region of Wasolon (where Mali abuts Guinee and Cote d'Ivoire), via overbroad generalisations on this basis about the representation of West Africa in colonial sources, to pronouncements about ethnology, anthropology and the state of the world that are most charitably interpreted as strategically provocative. Revisiting this work alongside the absorbing biographical collection that Amselle has co-edited from a conference held in 1996--seventy years after Maurice Delafosse's death--reinforces a sense that Amselle's opinions are invariably expressed in two modes. The initial version of an argument is cautious, nuanced and credible; its restatement is as consistently overstated and implausible (and therefore seized upon gratefully by his critics, to the detriment of his genuine originality).

The cautious parts of Mestizo Logics, to look at this first, closely follow or exemplify the argument of the essay with which Amselle introduced an earlier co-edited book (Amselle and M'Bokolo, 1985). Amselle's argument there counterpointed two apparently contrary ideas: that ethnies (ethnic groups) were false archaisms--invented under colonial rule--yet contemporary ethnies were like their pre-colonial predecessors in being heterogeneous in composition (1985: 23, 37). This counterpoint was more interesting than some of his critics, who seized merely on its first statement (attributing the creation of ethnies to colonisers), had allowed (Fardon, 1996: 132-6, commenting upon Schilders and Binsbergen, 1993: 7-8; Austen, 1992: 286). Amselle could be interpreted as suggesting that the historical continuity or discontinuity of ethnicity in Africa is a matter of perspective. It could equally be true that contemporary ethnicities (in the plural) have African historical precedents and that contemporary ethnicity (as a type of difference with characteristic presuppositions and idioms) is unlike its historical forebears. Everything depends upon the criteria in terms of which resemblance is sought, and how the instances (ethnicities) are heuristically distinguished from the idiom (ethnicity) to see how they articulate. What survives of this argument in Mestizo Logics has become surrounded by broader claims that are both less nuanced and less plausible.

The central chapters of Mestizo Logics derive from Amselle's long-term local researches and remain close to the spirit of the 1985 article. The argument runs something like this. Prior to colonisation there existed `chains' of societies and cultures between which boundaries were fluid and contextually variable (whether we look at them in terms of processes of ethnic or religious identification or in terms of political or economic organisation). In Wasolon these social and cultural forms would later be identified as Senufo or Minyanka (at the `savage' or pagan end of a continuum), Bambara (at its barbarian middle) and Fulani and Malinke at its Muslim pole (pp. 148-9). Relations between these categories of identification may be envisaged as a set of transformations that occurred with economic and politico-religious prodding. The systematic …