In his review article, Terence Ranger raises two major issues. First, he is concerned about the relation between exoticizing ideas about an ‘African occult’ in Western societies (for example, Scotland Yard investigations, media reports, etc.) and Africanist work on this topic. This is a pertinent issue. Pointing out worrisome overlaps between popular ideas about bloodthirsty ritual practices and scholarly research, Ranger urges us to reflect more deeply about the political field into which the knowledge we produce about ‘occult matters’ is being launched. Second, Ranger critiques a certain type of study – especially work on ‘occult economies’ and ‘the modernity of witchcraft’ – for inventing an ‘aggregated African occult’ that is too generalizing and present-centred to achieve real insight into the modes through which different African societies grapple with questions of evil. Instead, he advocates a historically grounded, ethnographically specific perspective. Both issues raised by Ranger are central to Africanist scholarship, and need our utmost attention. Therefore it is laudable that Africa is creating space for further debate.

There is a tension in Ranger's argument. For even the position of ‘splitter’ rather than ‘lumper’ (2007: 277) is predicated on acknowledging a broader category such as that of the ‘occult’ (a notion which Ranger clearly uses with unease). This tension can in my view not be resolved by retreating into the study of the particular, but needs to be acknowledged as intrinsic to our scholarship. Even though I have great sympathy for detailed studies (and if pressed would certainly side with the ‘splitters’), I consider it unproductive to play off the level of the particular (or local) against that of the general (or global). At least in my understanding, the notion of ‘occult economy’ and the framework of the ‘modernity of witchcraft’ (interestingly, Ranger as well as ter Haar and Ellis, notwithstanding major differences, converge in their critique of the Comaroffs) may well be employed for the sake of detailed research. Highly diverse societies may still face similar challenges. As I have also tried to show in my own work, attention paid to modernity and globalization does not necessarily imply a disregard for local specificities, but may, on the contrary, entice a historical and ethnographic study of how the aggregation of the occult occurs in particular settings. Examples that come to my mind are missionary demonizations of local religious traditions, or Nigerian films of the Nollywood type that excel in visualizing witchcraft, revenge ghosts, ritual murder and the like, and which certainly enhance prejudices about Nigerians as ‘occultist’ throughout Africa. Remarkable figures such as Credo Mutwa – ‘this old charlatan’ (ibid.: 274) – also partake in practices of aggregating an African occult, suggesting spectacular links between Zulu visionary practices and Hollywood movies such as Spielberg's ET. In my view, such phenomena require more than unmasking them as inauthentic. They call us to pay detailed attention to actual practices of aggregation that mobilize resources from far away and close by. It goes without saying that such research requires historical and ethnographic specificity. At the same time we need to keep on reflecting on the very notions and concepts we employ to make sense of what we find ‘on the ground’.

How far is the proposition made by Gerrie ter Haar and Stephen Ellis useful in this endeavour? Their main point is that Ranger, notwithstanding his plea for a more specific approach, still mobilizes a broad notion of ‘the occult’, and thus takes part in the very project he critiques. Explaining that they do not use the notion of the ‘occult’ in their own work at all (hence they feel misinterpreted by Ranger, who charges them with being party to the project of ‘aggregating the occult’), these authors advocate discarding this notion altogether. As an alternative, they introduce religion as a more ‘neutral’, ‘value free’ and more encompassing term (p. 400). They argue that a Christian, moralistic understanding of religion as ‘whatever is good and life-affirming’ underpins the work of Ranger and other Africanists. This needs to be replaced by a focus on spiritual powers that act both constructively and destructively, depending on context. This shift, they argue, would allow scholars to get beyond a dualist opposition of the occult (understood as evil) and religion (understood as devoted to the moral good) that informs current work on ‘the occult’.

I find their plea to critically rethink the distinction between ‘the occult’ and ‘religion’ (which obviously resonates with older debates about the question of whether magic belongs to religion or not) important on empirical and conceptual grounds. Their piece offers much for further discussion. Nevertheless I remain to be convinced that adopting a definition of religion that incorporates what other scholars frame as ‘occult’ would offer a viable solution with regard to the two issues that, as outlined above, stand central in Ranger's piece, which forms the immediate target of the rejoinder by ter Haar and Ellis.

Regarding the first issue, the problematic convergence between popular and scholarly representations, ter Haar and Ellis offer no solution. While it is true that such things as ritual murder actually occur, the question is still how to frame our reports on such matters. To state that this is part of African religion still affirms a problematic exoticizing view. As I see it, certainly with the increasing presence of Africans in our religiously and culturally plural societies, we need a better grasp of the social field in which Africanist works intersect with bureaucratic institutions and policy makers, migrant communities, politicians and priests in Africa, and popular representations of witchcraft, magic and voodoo. This is key for finding a mode of participating in public debates and dialogues with policy makers and official institutions that questions stereotypes without neglecting crime.

With regard to the question of how we as scholars frame ‘occult’ matters – the second major issue raised by Ranger – here again I wonder how far adopting a broad definition of religion will get us. There is an ongoing debate in anthropology and religious studies about the very problem of defining religion (in fact, its ‘definability’). While ter Haar and Ellis make use of this work (for example, studies by Asad and Masuzawa) in critiquing the definition of religion that underpins the work of Ranger, they still come up with their own definition that takes as a departure point ‘belief in the existence of an invisible world’ (p. 2). The resonance with earlier definitions, from Tylor to Geertz, is obvious and one wonders what is new. Definitions per se cannot resolve intricate scholarly issues. Instead of following ter Haar and Ellis in adopting an allegedly neutral, value-free definition of religion in terms of belief, I plead for a thorough, self-reflexive study of the use of the notion of religion on the ground, as well as in African scholarship. There are many examples that show that the category of religion itself operates in a politics of inclusion and exclusion and thus is never neutral. As pointed out by David Chidester in Savage Systems, in the nineteenth century ‘religion’ featured as a category of exclusion: Africans did not have it, and hence needed to get it from more enlightened outsiders. In the meantime, Africans have been characterized, equally problematically, as ‘incurably religious’. Many Africanists like myself are familiar with Christians subsuming magic, juju, witchcraft, and other destructive spiritual forces under the Christian category of evil, whilst advocates of the respectability of African Traditional Religion (writ large) insist that such destructive, ‘occult’ practices would not at all be part of religion, charging Christians with misrepresenting African religion. Lumping a diverse set of phenomena together as ‘religion’ obfuscates a clear grasp of the use of such terms as ‘religion’, ‘witchcraft’ and other spiritual matters on the ground (as also pointed out by Ranger). The actual use and contestation of such terms itself requires detailed study.

Ter Haar and Ellis lack reflexivity regarding their own definition, which still echoes a now much-critiqued understanding of religion in terms of belief. While there are good reasons to be critical about contemporary scholarship on the ‘occult’, it needs to be acknowledged that one of its big merits is the focus on intersections between the spirit world and various social domains, including politics, economics, law, or technology. This is the strong point in the work on witchcraft by, for example, Peter Geschiere, Isaak Niehaus, Harry West or Adam Ashforth. Containing spiritual power, called upon for whatever purpose, within the category of religion narrowly understood as belief, limits the scope of our inquiry and ultimately contradicts one of the central concerns of the work of ter Haar and Ellis: pointing out how invisible spirits are entangled with all domains of life.